Li V

HARVARD UNIVERSITY

LIBRARY

OF THE
MUSEUM OF COMPARATIVE ZOOLOGY

GIFT OF

Linnean Soc. London.
1831-1889

THE

TRANSACTIONS

**-£

OF

THE LINNEAN SOCIETY

OF

LONDON.

SECOND SERIES— VOLUME II.

ZOOLOGY.

LONDON:

PRINTED BY TAYLOR AND FRANCIS, RED LION COURT, FLEET STREET:

SOLD AT THE SOCIETY'S APARTMENTS, BURLINGTON-HOUSE.
AND BY LONGMANS, GREEN, AND CO., PATERNOSTER -ROW.

1879-1888.

S-E.S-L

CONTENTS.

PART I.— December, 1879.

I On the Genus Actinometra, Mull., with a Morphological Account of a New Species
(A. polymorph*) from the Philippine Islands. -Tart I. By P. Herbert Car-
penter MA, Assistant Master at Eton College. {Communicated by W. B.
Carpenter, C.B., M.D., BB.B., F.B.S., F.B.S.) (Plates I.-VIIL). . page 1

PART IL— March, 1881.

II. On some New Species of Nudibranchiate Mollusca from the Eastern Seas. By

Cuthbert Collingwood, M.A., M.B., F.B.S., 8fc. (Plates IX. & X. and Wood-
cut.)

III. On the Anatomy of Ants. By Sir John Lubbock, Bart., M.P., F.B.S. , F.L.S.,

B.C.B., LL.U., rice-Chancellor of the University of London, President of the
Entomological Society. (Plates XL & XII.) 141

IV . On the Extinct Band- Tortoises of Mauritius and Rodriguez. By Alfred C. Haddon,

B.A., Scholar of Christ's College, and Curator in the Museum of Zoology and
Comparative Anatomy of the University of Cambridge. {Communicated by Prof.
Newton, M.A., F.B.S.) (Plate XIII.) 155

PART III.— February, 1S82.

V. On the Morphology of the Skull in the Amphibia Urodelia. By Prof W. Kitchen

Parker, F.B.S., F.L.S. (Plates XIV.-XXI.) 165

VI. On the Tusks of the Fossil Walrus, found in the Bed Crag of Suffolk. By E. Ray

Lankester, M.A., F.B.S., F.B.S., Professor of Zoology and Comparative Anatomy
University College, London. (Plate XXII. and Woodcut.) 213

in

iv

J

PART IV.— March, 1882.

VII. The Parasites of Elephants. By T. Spencer Cobbold, M.D., F.R.S , F.L.S., fyc.
(Plates XXIII. & XXIV.) page 223

PART V.— July, 1882.

VIII. On the Digastric Muscle, its Modifications and Functions. By G. E. Dobson,
M.A., M.B., F.L.S. (Plate XXV.) 259

PART VI.— April, 1883.

[X. On the Clasping-Organs Ancillary to Generation in Certain Groups of the Lepi-
doptera. By Philip Henry Gosse, F.R.S. {Communicated by R. M'Lachlan,
F.R.S., F.L.S.) (Plates XXVI.-XXXIII.) 265

PART VII.— September, 1883.

X. On certain Points in the Anatomy of the Polynoina, and on the Polynoe (Lepidonotus,
Leach) clava of Montagu. By Alfred Gibbs Bourne, B.Sc. Loud., Univ.
Scholar in Zoology, and Assist, in the Zoological Laboratory, University College,
Loudon. (Communicated by Prof. E. Ray Lankester, M.A., F.L.S.) (Plates
XXXIV-XXXVI.) 347

PART VIII.— August, 1883.

XI. On Simondsia paradoxa and on its Probable Affinity icith Sphserularia bombi. By
T. Spencer Cobbold, M.lJ., F.R.S., F.L.S., Correspondent of the Academy of
Sciences of Philadelphia. (Plate XXXVII.) 357

PART IX.— October, 1883.

XII. On the Testis of Limulus. By W. B. S. Benham, Esq. (Communicated by Prof.
E. Ray Lankester, M.A., F.L.S.) (Plate XXXVIII.) 3G3

PART X.— April, 1884.

XIII. The Metamorphosis of Filaria sanguinis bominis in the Mosquito. By Patrick
Manson, M.I)., Hong Kong. (Communicated by Br. Cobbold, F.R.S., F.L.S.)
(Plate XXXIX.) 307

[ v ]

PART XI.— December, 1884.

XIV. On the Compound Vision and the Morphology of the Eye in Insects. By B.
Thompson Lowne, F.B.C.S., F.L.S., Lecturer on Physiology, Middlesex Hospital,
formerly Arris and Gale Lecturer, Royal College of Surgeons. (Plates XL.-

XLIII.) Pa§e 389

PART XII.— November, 1885.

XV. Contributions to the Knowledge of the Genus Anaphe, Walker. By Lord Wal-

singham, M.A., F.L.S. (Plates XLIV. & XLV.) 421

PART XIII.— August, 1884.

XVI. On a new Species of Coelacanthus (C. Tingleyensis)/ro?ra the Yorkshire Cannel Coal.
5y James W.Davis, M.S., .F.G.S. (Plates XLV I. -XLIX.) 427

PART XIV.— January, 1885.

XVII. On three new Species of Metacrinus. By P. Herbert Carpenter, D.Sc,
Assistant-Master at Eton College. With a Note on a new Myzostoma, by Prof
L. von Graff, Ph.D. (Communicated by Dr. W. B. Carpenter, F.B.S., F.L.S.)
(Plates L.-LII.) 435

PART XV.— August, 1885.

XVIII. On the Breeding of Salmon from Parents which hare never descended to the Sea.
By Francis Day, F.L.S. (Plates LIU. & LIV.) 447

PART XVI.— October, 1885.

XIX. Golfingia Maclntosliii, a new Sipuncu lid from the Coast of Scotland. By~E. Ray
Lankester, M.A., LL.D., F.B.S., Jodrell Professor of Zoology in University
College, London, Fellow of Exeter College, Oxford. (Plates LV. & LVI.) . 469

PART XVII. — February, 1886.

XX. On the Variations in the Form of the Cirri in certain Comatula?. By P. Herbert

Carpenter, D.Sc, F.R.S., Assistant Master at Eton College. (Communicated by
W. Percy Sladen, Sec. Linn. Soc). (Plate LVII.) -475

PAP^T XVIII.— November, 1888.
Title-page, Contents, and Index.

second series. — zoology, vol. II.

TRANSACTIONS

OF

THE LINNEAN SOCIETY.

I. On the Genus Actinornetra, Mull., with a Morphological Account of a new Species
(A.) polymorpha from the Philippine Islands. — Parti. By P. Herbert Carpenter,
M.A., Assistant Master at Eton College. {Communicated byW. B. Carpenter,
C.B., M.B., LL.D., F.B.S., F.Z.S.)

(Plates I.-VIII.)

Read June 21st, 1877.

I HE principal part of the investigations, of which the results are set forth in the fol-
lowing pages, were carried on during the year 1876, in the Zool.-Zootomical Institute of
the University of Wiirzburg, under the superintendence of its director, Prof. C. Semper.
I would here express my most sincere thanks to Prof. Semper, not only for the
generous manner in which he placed at my disposal all his specimens of Act. polymorpha,
which he had himself collected in the Philippine Islands, hut also for the ready and
valuable advice which he constantly afforded me during the progress of my work, and
for the free use which he permitted me to make of his extensive library. I am also
greatly indebted to Dr. Sandberger, Professor of Geology in the University of vYiirzburg,
to the authorities of the University Library, and to Professor Dr. Halm, of the Royal
Library at Munich, for the means of reference to many books which would otberwise
have remained inaccessible to me ; and I desire to record my best thanks to all these
gentlemen for the ready kindness with which they met my frequent and numerous
wants.

I. Historical.
(§ 1) In the remarkable work of Linckius 1 upon the " Sea Stars," the modern family
Comatulidce (J. Miiller) is described under the name of Crinita), or Comatse Stelloe, as a
group distinct not only from the Asterida?, but also from the Ophiuridaa and from
Astrophyton {Euryale, Lamarck), with which they have been united by many later
systematists.

1 Johannis Henkici Lisckii Lipsiensis ' De Stellis Harinis liter singularis,' Lipsise, 1733, p. 53.
SECOND SERIES. — ZOOLOGY, VOL. II. 1

2 ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA.

Liuck included three genera in this family, or, as he called it, " Classis." The first of
these he named Ae/caKvij/ioq, to indicate " stellam marinam decern caudis crinitis radi-
antem ; " and he referred to it three species : — (1) The " Crocea zaffarcma Neapolitanorum"
or 8e/caSa<TuaKT(>'oei£)1c of Fabius Columna x, whose description he quotes ; (2) the Decempeda
Cornubiensium of Llliuyd 3, Avhich Linck figured and named " Stella Se/ca/ci')),uoc rosacea ; "
and (3) the " Ae«:aKv»j/uoc Jtmbriata Barrelieri " 3, which was named by Liuck barbata, as
he supposed it to be different from the other two. All three, however, are really identical,
being simply local -varieties of one and the same species, viz. the British Antedon rosacea,
or the Comatula mediterranea of Lamarck. Thus, Fabius Columna described the
Neapolitan variety, and Barrelier another obtained at the mouth of the Tiber, while Llliuyd
based his description upon specimens found upon the coast of Cornwall near Penzance.

Linck's second genus, the T pianaileKUKv^oc, was based upon a specimen with thirteen
arms, previously described by Petiver4 as " Stella chinensis ;" this specimen, however,
was suspected by Linck to have been mutilated. His third genus he called " Caput-
Medusce," and described it as including those specimens which " ex centro corporis parvi
umbonatique per quiuque truncos primum bifidi, mox nullo constanti nurnero multifidi,
in 60 et plurcs surculos geniculatos rectos simpliccs abeunt, quos gracilescentes fibrilkc
arise pilorum instar vestiunt."

Linck referred two species to this genus, viz. Caput Medusae cinereum and C. brnnnum;
and he gave good figures of both (tab. xxi. n. 33, and tab. xxii. n. 31), from which it
may be determined with tolerable certainty that they represent species now known to
belong to two different types among the Coinatuke — namely, to the genera Antedon
and Actinometra respectively.

(§2) Although Llliuyd5, and after him Bosinus1', had explicitly pointed out the
relationship between the recent Comatulae and the fossil Crinoidea, and although Linck,
while supporting and repeating Llhuyd's views, had clearly differentiated the former
from the Asteroidca and Ophiuroidea, yet Linnaeus 7, instead of adopting the more
correct views of some of his predecessors as to the true relations of the Crinoidea, was
so misled by the jointed structure of their stems as to rank them among zoophytes in
his genus Isis, whilst he grouped the Comatulidas, together with all the other Starfish,
under one common name Asterias. Linck's three species of Decacnemus were rightly
regarded by him as identical ; and he placed them, together with Petiver's Stella chinensis,
in one species, Asterias peclinata, to which he also referred a specimen previously
described by Betzius 8. We now know, however, that this last is an Actinometra, dif-

1 Phytobasanus, sivc Planfcarum aliquot Historia. Neapoli, 1592.

- Eduaedi Luidi ' Lithophylacii Britanniei Icknographia ' p. 149. Londini, 1699.

'■' Jac ib] Bakeelieei ' Plantae per Galliam, Hispaniain ct Italiam observatrc ' Paris, 1714, p. 131.

4 ' Gazophylacium Natura: et Artis,' Londini, 1711 ; and also ' Aquatilium animalium Aniboincnsiurn Ioones ct
Nomina,' 1713.

5 Prsclcctio de Stellis marinis Oceani Brit, nee non de Asteriarum, Entrocliorum, et Encrinorum Origine, pp. 149-
155, Oxford, 1733.

c Tontaminis do Litliozois ae Lithopbytis olini marinis, jam vero subterraneis, prodromus ; sive de stellis marinis
quondam, nunc fossilibus, disquisitio. Hamburg, 1719.

7 'Cystoma Natural,' editio decima tertia (Lipsire, 1783), pars vi. p. 3100.
s Nova Acta, Stockholm, 17S3, p. 234, n. 12.

ME. P. II. CAEPENTEE ON THE GENUS ACTLXOMETEA. 3

fering very considerably from the type represented by Decacnemus (Antedon) rosacea,
although resembling it in only having ten arms.

In like manner both of Linck's species of his genus Cajiut-HIedusa?, the one an
Antedon and the other an Actinometra, were united by Linnaeus, together with a many-
armed specimen described by Retzius, into one species, Asterias maltiradiata.

(§3) For some years after the commencement of the present century the Linnean
nomenclature held its own, and the few species of recent Comatukc with which the
naturalists of that time were acquainted were described as different species of the
Linnean genus Asterias.

The first among the post-Linnean zoologists who recognized the claims of this form of
Sea Star to a distinct generic rank was De Ereminville \ who in 1811 presented to the
Societe Philomatique de Paris a " Memoire sur un nouveau genre de Zoophytes de
l'Ordre des lladiaires," to which he gave the name of Antedon. His definition of the
genus was as follows : — " Animal libre, a corps disco'ide calcaire en dessus, gelatineux
en dessous, environne de deux ran gees de rayons articules, pierreux, perces dans leur
largeur d'un trou central ; ceux du rang supcrieur plus courts, simples, et d'egale
grosseur dans toute leur longueur ; ceux du rang inferieur plus longs, allant en dimi-
nuant de la base a la pointe, et garnis dans toute leur longueur d'appendices alternes
egalement articules; bouche inferieure et centrale."

It is not very clear which of the two apertures on the ventral (or, as he called it,
inferior) side of the disk was regarded by De Freminville as the mouth ; it is very pro-
bable that, as he was only able to examine a spirit-specimen, he failed to recognize more
than one — that namely, which, placed at the extremity of a long tube projecting from
a point near the centre of the disk, we now know to be the anus.

Adams3, who had studied living specimens of Linck's Decacnemus rosacea, had, how-
ever, pointed out some years previously the existence of two orifices to the digestive
cavity ; but his observations seem to have escaped notice ; for Lamarck 3, Miller i, and
many other naturalists, all regarded the aperture at the end of the anal tube either as the
mouth alone or as a combined mouth and anus ; and it was not till 1823 that the exist-
ence of distinct oral and anal orifices was fully recognized.

It was announced as a new discovery by Leuckart 5, in a letter to Von Schlotheim ;
and he was followed shortly afterwards by Meckel9, Gray7, and Heusingers.

1 Nouv. Bull. d. Scien. par la Soc. Philomat. torn. ii. p. 340. Paris, 1811.

2 " Description of some Marine Animals found on the Coast of Wales," Trans. Linn. Soc. vol. v. p. 7 (1800).

3 Systeme dAnimaux sans Verfcebres, 2me e'd. (Paris, 1816), torn. ii. p. 532.

4 A Natural History of the Crinoidea (Bristol, 1821), p. 128.

5 Von Schlotheim, Naeht. z. Petrcfact. ALtli. ii. p. 48 (Gotha, 1S23); and Leuckart, "Einiges iiber Asteriden
Gesehlecht Comatula Lam. uberhaupt, und iiber Coi . a insbesondere," Zeitsch. fur organ. Physik, iii.
p. 385 (1833).

6 " Leber die Ocffnungen des Speisekanals bei den Comatulen," Meckel's Archiv fur Physiol. Band iii. p. 470
(1823).

7 " Notice on the Digestive Organs of the Genus Comatula and on the Crinoidea of Miller,'' Ann. of Philos. n. s.
vol. xii. p. 392 (1826).

8 " Bcmerk. iiber d. Verdauungskanal dcr Comatulen, Meckel's Archiv" fur Physiol. 182(5, p. 317; and "Anat. Uu-
tersuch. d. Comatula mediterranea," Zeitsch. fur organ. Physik, iii. p. 300 (1S:j3).

1*

4 ME. P. H. CAEPENTER ON THE GENUS ACTINOMETRA.

Dc Preniinville's specimen was found on the keel of a vessel which had come from a
warm climate ; it had ten arms and twenty cirrhi, and was named by him Antedon gor-
gonia. He gave no further description of it, but simply referred to the figure of Stella
decacnemus rosacea, Linck, in the ' Encyclopedie Methodique ' \ which represents the
ordinary European Comatula rosacea, as it is now called.

This species, however, is not identical with De Premiuville's Antedon gorgonia, which
was referred by Lamarck2 to his Comatula carmata, under which name he described
some specimens brought by Peron from the Isle de Prance. Nevertheless the two
species resemble one another in some important points, viz. the presence of ten arms,
of a central or subcentral mouth, and of an excentric anal tube.

In 1815 Leach3 rescued the three genera contained in Linck's classis CrinitcB
from the confusion of the Liunean genus Astcrias, and united them into one genus,
Alecto, comprising three species, viz. Alccto europcea (= Decacnemus rosacea, Linck),
Alecto horrida (= Caput-Medusoe, Linck, or Asterias multiradiata, Linn.), and Alecto
carinata (which seems to have been the same as De Premiuville's Antedon gorgonia).

Leach defined Alecto as having the " os inferius, irregulare," a description which would
suit equally well either for the true mouth or for the anal opening, though perhaps
more applicable to the former. He seems, however, like his predecessor De Preminville,
to have regarded the mouth as situated at the extremity of the anal tube ; for in the
explanation to Schweigger's figure1 of Leach's specimen of Alecto horrida the latter is
described as the " rohrenformig hervorstehender Mund." It is obvious, therefore, that
we cannot make any use for systematic purposes of the definitions of Antedon and Alecto
as given by Leach and De Preminville respectively, as far as the position of the mouth
is concerned.

Schweigger's figure of the disk of Alecto horrida shows clearly enough that the five
trunks of the ambulacral grooves converge towards the centre of the disk, as in Antedon
rosacea (Alecto europcea, Leach), Plate I. fig. 1. Leach's Alecto horrida was therefore
a true Antedon in the modern sense of the term, although belonging to that division of
the o-enus in which the repetition of the bifurcation of the ten primary arms is carried
to a great extent.

(§4) Leach was apparently unacquainted with the memoir of Preminville; but the
same was evidently not the case with Lamarck (1S1G), who, like Leach in the previous
year, united Linck's three genera into one, to which he gave the very appropriate name
Comatula5. His definition of the genus differs but little from that given for Antedon
five years previously by De Preminville, whose original specimen Lamarck seems to have
examined ; and it is difficult to see why he did not adopt the name Antedon to designate
the genus, which, like Leach and De Preminville, he clearly distinguished as belonging
to a different type from the Asteridse, Ophiuridoe, and Eur y ale.

1 Partie des Vers, pi. 42. fig. G. 3 Tom at. p. 534.

3 Zoological Miscellauies, vol. ii. p. 61 : London, 1815.

4 Beobacht. auf naturhistorischen Reiseu, p. GO, and Taf. iv. fig. 42 : Berlin, 1819.

5 Turn. cit. p. 530.

ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA. 5

Laniarck included eight species in bis new genus Comatula. In only three of these is
the mouth at or near the centre of the disk, viz. C. mediterranea (= Stella decacnemus
rosacea, Linck), C. carlnata(= Antedon gorgonia, T?rem. ?), and C. adeonce.

In all the other five species described by Lamarck, viz. C. Solaris, C. brachiolata,
C. rotalaria, C. fimbrlata (= Stella chinensis, Petiver), and C. multiradiata (= Caput-
Medusce brunnum, Linck), the mouth (as I know from examination of the collection of
Comatula? in the Paris Museum, which still contains many of Lamarck's original spe-
cimens) is nowhere near the centre of the disk, which is occupied by the anal tube, but
is excentric, or even marginal. At the same time the five primary groove-trunks do not
converge towards the centre of the disk, as in Antedon rosacea (Com. mediterranea, Lam.)
(PL I. fig. 1), but they unite more or less completely into a horseshoe-shaped furrow, at
one point of which is situated the excentric mouth (PL I. figs. 2-5).

This type will be described further on, under the name of Actinomelra. Lamarck, who
found it in more than half of the species constituted by him, seems to have regarded
it as common to all ComatuLw His description of it is worth quotiug, as it is the first
notice of a true Actinometra that I have been able to find. lie says 1 : — "Le disque
iuferieur ou ventral offre un plateau orbiculahv plus large que le dorsal, entoure de rayons
simples, cirreux. Pres de la circonference de ce plateau, on apercoit un sillon irregu-
lierement circulaire qui s'ouvre sur la base des rayons pinnes, et se propage le long de
leur face inferieure, aussi que de celle des pinnules. Cc sillon neanmoins, ne s'approche
point de la bouche [i. e. the anal tube] et ne vient point s'y reunir, comme cela a lieu
pour la gouttiere des rayons dans les Asteries. Au centre du disque iuferieur ou ventral
des Comatules la bouche, membraneuse, tubuleuse, ou en forme de sac, fait une saillie
plus ou moins considerable suivant les especes."

Although Antedon and Alecto were both constituted previously to Comatula, yet
Lamarck's authority was sufficient to establish the latter name, and to bring it into
general use, though Cuvier adopted Leach's genus Alecto, and used it in preference to
Comatula. The latter, however, was more generally employed by all subsequent
observers (who pointed out Lamarck's error respecting the position of the mouth)
thenceforward till the time of Johannes Miiller.

(§ 5) During this period the skeleton both of the recent and of the fossil Crinoids was
made the subject of careful investigations by Miller3 and Goldfuss3. The latter
author divided his class Stellerites into two Orders : — (a) Stilasteritse, or Gestielte
Seesterne ; and (b) Asterites liberi, or Preie Seesterne. The former he again divided into
Articulata and Inarticulata, placing in the first group a number of fossil Mesozoic
Crinoids, viz. Eugeniacvinites, Miller, Solanocrinites, Goldf., Pentacrinites, Encrinites,
and Ajuocrinites.

In his descriptions of these genera he adopted and considerably improved the some-
what inapplicable system of nomenclature, introduced by Miller for the parts of the
skeleton of the fossil Crinoids and of Comatula.

Portions of his diagnoses of Eugeniacrinites and Solanocrinites are of considerable
interest, both zoologically and morphologically. Of the former, he says 4 : —

1 Tom. tit. p. 532. '; Loc. cit. 3 Petrefacta Germanic, i. (Dusseldorf, 1826-35). 4 Turn. tit. p. 102.

6 ME. P. H. CAEPENTEE ON THE GENUS ACTIXOITETEA.

" Die kurze runde mit eiuem runden Kanale durchbohrte Saule besteht aus weuigen
walzigen verlangerten Gliedern, und nininit am obern Ende allmalig zu. Hire Gliede-
rung wird oft nur durcb Binge angedeutet, unten endet sie in starken Wurzelu. Das
letzte verdickte Saulenglied vertritt die Stellc des Beckens und articulirt durcb eine
Gelenkflache mit den Bippengliedern."

Tbe term " Becken " is bere meant to signify the circlet, of basals which, in Penta-
crinus and other stalked forms, intervenes between tbe stem and tbe circlet of first
radials, tbe " Bippenglieder " of Goldfuss.

These basals are also present in Solanocrinites, which genus, as Goldfuss well remarks,
" hat in verscbiedener Hinsicbt Aehnlichkeit mit den Pentacriniten, und bildet zugleich
einen Uebergang zur Gattung Comatula."

Its most important characters are as follows ! : —

" Die Saule ist sehr kurz, beiuabe so dick als der Kelcb, funfseitig und an ihrer
Basis nicbt mit Wurzelsprossen, sondern mit ausstrablenden B/unzeln versehen. Hire
Glieder sind mit einander verwachsen und haben an den Seitenfiacben Gelenkhohlungen
fur den Ansatz zahlreicher dicker Hiilfsarme. Die oberste Gelenkfliicbe der Saule zeigt
fiinf strablenformige Erhabenheiteii, auf welchen das Becken articulirt. Das Becken
wird nicht durcb das oberste Saulenglied vertreten, sondcrn es besteht aus fiinf Gliedern,
welche zwischen die Nahte der fiiuf Bippenglieder eingefiigt sind, oder sie bedecken.
Die fiinf Glieder des Beckens bilden entweder nur schmale Strahlen, die zwischen die
Nahte der Bippenglieder einsenken (S. costatus, S. scrobiculatus), oder sie sind breiter,
stossen seitlicb an einander, und stellen eine tiefe mit 5 Strahlenfurclien ausgehohlte
Gelenkflache dar (S.jcegeri)."

The first genus in Goldfuss's order " Asterites liberi" is tbe Comatula of Lamarck,
of which he says 2, '; Diese Gattung bildet den Uebergang von den Stilasteriten zu den
freien Seesternen und steht zunachst mit den Solanocriniten in nachster Verwandtschaft."

Besides describing five fossil species, he gives some account of two recent ones,
dwelling more especially upon tbe structure and composition of the skeleton.

Thus "Bei der G. meditcrranea besteht die Saule aus drei Gliedern : das Becken fehlt,
und die Bippenglieder sitzen unmittelbar auf dem letzten Saulenglicde." "Bei der in
den Ostindiscben Meeren lebenclen C. multiradiata Lam. bingegen, finden sich Becken-
glieder, so dass man berechtigt sein konnte sie als eine eigene Gattung zu betracbten.
Ihr Saulenrudiment besteht aus einem einzigen schiisselformigen Gliede, an dessen
Bande fiinf schmale dreieckige Beckenglieder ansitzen und mit ihm verwachsen sind.
Diese stossen mit ihren Seitenfiachen nicbt an einander, sondern stehen so weit entfernt
dass die ersten Bippenglieder unmittelbar zwischen ibnen auf dem Siiulengliede ansitzen,
und sie durch einen Abscbnitt der unteren Ecke zwischen sich aufnehmen. In der
Mitte des innern unteren Ptandes jedes Beckengliedes entspringt ein zahnformiger
Eortsatz, der sich als knospeliger schmaler Streifen bis zum Mittelpunckte des Sau-
lengliedes verlangert, in welcbem er durch eine Binne aufgenommen ist. Diese Becken-
glieder sind also denen der Solanocriniten vollkommen analog."

(§ 6) Goldfuss, who, though acquainted with the name of Alecto, yet used Comatula

1 Tom. tit. p. 166. - Tom. at. p. 202.

ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA. 7

in preference to it, paid very little attention to the soft parts of either of the Comatula?
which he dissected.

In the following1 year, however, De Blainville ' described the visceral mass at some
length. Like his predecessors, he adopted Lamarck's genus Comatula, making it the
only representative of his section, " Les Asterencrinides libres," while at the same time
he acknowledged the prior claims of de Ereminville's name Antedon. He was, of course,
acquainted with Lamarck's error respecting the position of the mouth, which he
described as " assez anterieure, isolee, membraneuse, au fond d'une etoile formee par cinq
sillons bil'urquc's." The species which lie dissected was a foreign one preserved in
spirit ; it had a large number of arms ; and from the not very clear description which he
gives of its ventral surface it would seem to have been a true Actinometra.

After speaking of the tentacular furrows on the ventral surface of the arms, he says 2,
" En suivant ces especes de sillons dont lc nombre est proportionnel a celui des digitations
du rayon, on arrive par un sillon unique pour chacun d'eux et qui en occupe la base, au
centre d'une sorte d'etoile a bords epais, franges, et par suite a la bouche qui est au fond.
L'etoile formee par la reunion des sillons des rayons n'est pas symetrique, c'est a, dire que
ses branches sont tres-inegales : les unes que j'appellerai les anterieures, etant bien plus
courtes que les autres, ou posterieures. II en est resulte que la bouche n'est pas au
centre de l'etoile, mais bien plus proched'un cote quede 1' autre : elle est assez difficile a.
voir au contraire d'un autre orifice, dont il va etre question, et que M. de Lamarck
paroit avoir pris pour elle. Elle est profondement enfoncee dans l'etoile des sillons : elle
est ronde, sans aucune armature et conduit immediatement dans l'estomac."

The above description implies, if I rightly understand it, that the mouth of De Blain-
ville's specimen was nearer to one side of the disk than to the other, so that the primary
trunks of the ambulacral grooves were of unequal lengths. This will be subsequently
seen (section 14) to be the principal distinctive character of the genus Actinometra.

De Blainville evidently attached no importance to the position of the mouth as a
character of systematic value in the determination of the species of recent Comatulce ;
and from his definition of it as " assez anterieure," it would almost appear as if he
supposed the other species to agree in this respect with the one dissected by him.

This is, in fact, the case in five out of the eight species described by Lamarck, with
which De Blainville was probably acquainted, and to which he added no new ones,
except that he gave the name of Comatula barbata to Linck's third species of Decaciiemus,
the fimbriata of Barrelier, or " barbata " of Linck. Lamarck had been uncertain to which
of his species he should refer it, although, as we have seen above (section 1), it is really
only a local variety of his C. medilerranea.

Like the other naturalists of his time (1S38), Agassiz 3 also adopted Comatula in preference
to the other generic names of this type, but defined it as having the " bouche centrale en-
foncee," and with the five " rayons du disque bifurques," thus limiting the number of
arms in the genus Comatula to ten only. At the same time he erected Lamarck's species
C. multiradiata, with sixty or more arms, into a new genus, Comaster, which he defined as

Manuel d'Aotinologie, (Paris, 1834) p. 2 19. : Op. cit, p. 251.

" Prodrome d'une Monographic des Eadiaires ou Echinodernies," Ann. des Scicu. Nat. 2e serie, Zool. vii. p. -~>7.

8 ME. P. H. CAEPENTER ON THE GENUS ACTINOMETEA.

follows : — " Ce genre a la meme organisation que le precedent ; mais les especes ont les
bras ramifies au lieu de les avoir simplement fourchus." Agassiz consequently used
Comatula simply as equivalent to the Decacnemus of Linck, while his new genus Comaster
was Linck's Caput-Mediisce, or the Comatula multiradiata of Lamarck. Of the seven
other species constituted by the last-mentioned naturalist, only two, C. rotatoria and
C. fimbriata, have more than ten arms ; in both of these the number of arms is usually
twenty, though it may reach twenty-four, or possibly even more. Strictly speaking,
therefore, these two species, according to the above definition, should be referred to
Comaster, and not to Comatula.

This character, the number of arms, upon which Agassiz founded a generic distinction,
is, in fact, extremely variable, and by no means of generic importance ; in fact, as
Goldfuss1 remarked a little later, " Wollte man mit Agassiz die Theilung der Arme
als hinreichendes Gattungsmerkmal ansehen, so wiirde man folgerecht gezwungen sein
fast jede Art der Crinoideen als Gattung aufzustellen."

Leach and Lamarck had already recognized this fact in uniting Linck's three genera
under a common name ; and it is not a little strange that Agassiz should have seen fit
to separate them again. His doing so, however, led to somewhat important consequences
from a systematic point of view. Turning to the fossil Comatula;, we find that Agassiz
erected the C. pinnata of Goldfuss into a new genus, Pterocoma, and grouped together
his other three species, C. tenella, C. pectinata, and C. filiformis, under the generic
name Saccoma ; while he expressed his belief thaf Solanocrinus was really related to
the Comatula;, and more especially to the problematical fossil described by Goldfuss
under the name of Glenotremites, which he rightly recognized as the centrodorsal piece
of a free-living Crinoid.

In the year 1810 a new fossil Comatula was described by Hagenow 2 under the name
of Hertha mijstica. The specimen, consisting of the united first radials and hemispherical
centrodorsal piece, was somewhat worn ; but Hagenow was able to recognize the resem-
blance between it and the remains of Solanocrinus, and the corresponding parts of
Goldfuss's Comatula multiradiata, except that he was unable to find any trace of the
external basals which Goldfuss had described in both the above cases ; and though he
seems, and (as we now know) correctly, to have suspected "das Vorhandensein etwa
verdeckt-liegender Beckenglieder," he was, of course, unable to come to any satisfactory
conclusion upon the point.

(§7) The year 18-10 is a noteworthy one in the history of our knowledge of the Crinoidea J
for it marked the appearance of the first of a series of classical memoirs by Johannes
Miiller, who laid the foundation of nearly all our knowledge of the zoology and morpho-
logy of the group : the first 3 of these was devoted to an anatomical account of the recent
and very rare genus Tentacrinus, together with many observations upon Comatula.

1 "Beitr. z. rctret'actenkunde," N. Acta Acad. Leop.-Carol. Nat.-Cur. xix.A. p. 348.

2 " Monogr. d. Riigen'schen Kreide-Versteinerungen, II. Abtheil. Eadiarien und Annulaten," Ncues Jahrb. Minc-
ralogie, 1840, p. G64.

3 " Uebcr den Bau des Pentacrinus ccvput-Medusce," Abhandl. d. Berlin. Akad. 1S43 ; Abstract in Monatsb. der-
selben, 1840 ; also in Wiegmann's Archiv f. Xaturgescb. 1840, i. p. 307.

ME. P. H. CAEPENTEE ON THE GENES ACTINOMETEA. 9

Miiller seems at first, not unnaturally, to have supposed that Goldfuss was right in
referring the rnany-armed specimen dissected by him to the Comatula multiradiata of
Lamarck, for which species he adopted Agassiz's name Comaster '. But he did not use it
precisely in the same sense as Agassiz, who, in his definition of the genus, makes no
mention of the external basals, the presence of wbicli was regarded by Goldfuss as the
principal character distinguishing Comaster from Comatula.

Miiller adopted Comaster 2 in the sense in which Goldfuss used the name ; and when
he subsequently discovered 3 that the Comatula multiradiata or Comaster of Goldfuss was
not specifically identical with the specimen described asComatitla multiradiata by Lamarck,
he retained the name Comaster for Goldfuss's specimen only, which, like Solanocrinus, is
remarkable for having " kleine basalia zwischen den Insertionen der Kelchradien, oder
sogennanten Beckenstucke welche den eigentlichen Comatulen ganzlich fehlen "4.

At the same time he gave a careful description 5 of Lamarck's original specimen of
Comatula multiradiata, based upon an examination of it by Troschel ; but as he regarded
Comaster and Solanocrinus only as one subgenus of Comatula, he gave it a new specific
name " multifula," on the ground that " die Comatula multiradiata Goldfuss, als die
zuerst genau beschriebene, den Speciesnamen multiradiata behalten muss." Lamarck's
specimen Avas thus restored by Miiller to its previous position among the " Comatulen
im engern Sinne, nainlich Gattung Alecto, Leach (Comatula, Lamarck)," which he
grouped together with Comaster into one family, Comatulinm. The fossil Solanocrinus
was regarded by him as identical with the latter form, while he referred the Hertha
mystica of Hagenow, and Pterocoma, Ag. (C. pinnata, Goldf.), to Comatula or Alecto;
for at that time (1841) he used the two names indifferently, considering them (as, indeed,
they originally were) equivalent to one another.

Goldfuss put forward about the same time a somewhat similar classification.6 In a
subsequent abstract (with additions) of his ' Beitrage zur Petrefactenkunde,' [which had
been published two years previously (1839)] he speaks of the two species dissected by
him as "die Typen der zwei nachst bezeichneten Genera (Comatula, Comaster), welche
daher nebst den zwei zuletzt folgenden (Solanocrimtes, Gasterocoma) als Verzweigungen
des Lamarck'schen weiten Geschlechts Comatula zu betrachten sind." He did not, how-
ever, agree with Miiller in regarding Solanocrinites and Comaster as identical, partly,
apparently, because nothing was known of the arms of the former, and partly because of
the differences in the form of the " Knopf," or centrodorsal piece, which he called a
short stem — although, as Miiller showed, this is not a character of any generic value.

Although Goldfuss had at first supposed 7 that the basals were really absent in Coma-
tula mediterranea, and that the first radials therefore rested directly upon the top of the
centrodorsal piece, or, as he expressed it, on the last stem-segment, he seems subsequently
to have changed his opinion ; for in his definition 8 of the genus Comatula, given in 1839,

1 " Beitrage zur Petrefactenkunde," ho. cit. p. 349. " Wiegmann's Archiv, 1840, i. p. 309.

3 " Ueber die Gattungen und Arten der Comatulen," Wiegmann's Archiv, 1841, i. pp. 14o, 147.

4 "Bau des Pentacrinus," loc. cit. p. L'7. 5 Wiegmann's Archiv, 1841, i. p. 147.

6 Neues Jahrbuch fiir Mineralogie, 1841, p. 818.

7 Petrefacta Germanise, torn. cit. pp. 203, 204. 8 Beitrage, &c. loc. cit. p. 349.
SF.COND SERIES. — ZOOLOGY, VOL. II. 2

10 ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA.

he says, " Auf deniletzten Saiilengliede ruhen fiinf Beckeuglietler, und auf jedem derselben
em Bippen- ( = second radial) undein Scliulterglied (third or axillary radial), auf welchem
zwei einfache Arme eingelenkt sind," from which it is evident that he was wrongly led
to regard the first radials as representing the basals of Comaster and JPentacrinus. This
mistake is hardly a surprising one when we consider the remarkable metamorphosis
undergone by the embryonic or primitive basals, and their concealed condition in the
adult Coinatula mediterrcmea.

Midler, who examined a very large number of species of Comatula, never found one in
which the basals appeared externally, as described and figured by Goldfuss in Comaster,
and remarked ' : — " Daraus geht hervor, dass die Gegenwart wirklicher Basalia ohne Zer-
legung bei einer lebenden Comatule, auch dann, wenn sic wirklich solche besitzt, schwer
zu erkennen sein muss. Die Unterscheidung der Comaster uud Comatula wird daher
bei der Ordnung der lebenden Comatulen unpractisch." In fact he appears to have
given up the genus Comaster altogether ; for he adds in a note : — " Kurzlich habe ich
die einzige im Museum zu Bonn befindliche Comatula mmltiradiata (nicht das von Goldfuss
zerlegto Exemplar, wovon ich nichts mehr vorfand) untersucht. Ich habe daran nichts
von Beckenstiicken erkennen konnen. Die Gattung Comaster ist daher wohl zu un-
terdriicken." He seems finally2 to have thought that it might possibly be identical
with the C. Bennett I of the Ley den Museum. As, however, Comaster has not been seen
by any naturalist since the time of Goldfuss, its position must still remain in doubt.

(§8) Up to the time of Midler no one paid any attention, from a systematic point of
view, to the arrangement of the tentacular furrows on the ventral perisome of the disk of
Comatula ; but Lamarck and De Blainville had, as we have already seen, examined and
described, with more or less accuracy, a condition which we now know to differ very
considerably from that presented by the Decacnemus of Linck, or the Antedon of De
Freminville. Both these observers seem to have regarded the former condition as the
normal one, and as common to all Comatuhe. Midler, who does not seem to have been
acquainted with their descriptions (for he makes no mention of them), took up the subject
systematically, and soon discovered that, using the distribution of the tentacular furrows
as a basis of classification, he could distinguish two, as he thought, very distinct types
of the genus Comatula, which he named Alecto and Actinometra respectively. In his
earlier communications3 on the subject he described the ordinary Comatula and Penta-
crinus as having a central mouth and symmetrically distributed tentacular furrows;
i. e. the five main trunks formed by the union of the furrows of the five groups of arms
converge directly towards the centre of the disk, being separated by five "interpalmar "
areas, one of which, slightly larger tban the rest, is occupied by the anal tube, which is
therefore excentric in its position (PI. I. fig. 5, An.).

During his visit to Vienna in 1810 Midler had an opportunity of examining an un-

1 'Uebcrdie Gattung Comatula, Lam., und ihre Arten,' Separatabdruck aus den Abhandl. Berlin. Akad. lS49,p. 8.

a Ibid. p. 29.

3 " Ban des Pentacrinus" loe. cit. p. 47, and Wiegm. Arcbiv, 1S40, i. p. 311.

ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA. 11

usually large specimen of the Comatula Solaris of Lamarck — unfortunately, however,
only a dry one, which he found to differ so greatly from the other Comatulce then known
to him, that he described it under the name of Actlnometra imperialis \ " welche gene-
risch von andern durch die Bildung ihres Scheitels verschieden zu sein schien. Auf
dem Scheitel der mit hlumenartigen Kalkhlattchen bedeckt ist, ist keine Spur von den
Furcken zu sehen, die bei den Comatulen von den Arm en zum Munde fiihren. Auch ist
dort nichts vom Munde zu sehen. Die Mitte der Bauchseite nimmt eine Ptohre ein.
Die Arme haben die ventrale Furche der Comatulen, die Furchen der zehn Arme miinden
aber in gleichen Abstanden in eine die Scheibe am Rande umziehende Cirkelfurche.
Diese eigenthumliche Bildung Hesse sich durch eine unsymmetrischc Vergrosserung
desjenigen Intertentacularfeldes, worin die Afterrohre steht, iiber den ganzen Scheitel
unci auf Kosten der anderen Intertcntacularfelder erklaren, so dass der Mund aus der
Mitte des Scheitels ganz an die Seite zwischen je zwei Armen gerath ; es ist mir aber
nicht gelungen den Mund hier zu finden." (PL I. fig. 2.)

In a subsequent visit to Lund, Muller examined two dry specimens of Comatulce, which
had been described by Eetzius 2 many years previously under the names of Asterias pec-
tlnata 3 and Asterias multlradlata. These he found to belong to the same type as the
Vienna specimen, which he had already designated Actlnometra imperialis, and which
he supposed to be distinct from the true Comatula Solaris of Lamarck. When he visited
Paris, however, in 1844, he examined Lamarck's original specimen of this species, and
convinced himself of its specific identity with his Vienna Actlnometra. Consequently he
withdrew the specific name "imperialis," and described the type simply as Actlnometra
solaris\ Muller was unable to determine the position of the mouth in the dry specimens
of the Lund aud Vienna Museums on which he founded his new genus Actlnometra; but
subsequently he was able to examine many spirit-specimens both of his typical species,
Act. Solaris, and also of other " Comatulen von jener Anordnung der Purchen, sowohl

zehnarmige als vielarmige Siehe die beistehende Pigur von Comatula JFahl-

berghil6" (PL I. fig. 3). This last species he describes a few pages further on as Coma-
tula (Actlnometra) JVahlberghll.

He did not, however, appear to regard the position of the mouth as of any systematic
importance ; for he goes on to say : — " Der Mund ist bei der in Frage stehenden Ab-

1 Wiegm. Archiv, 1841, i. p. 141 ; and " Neuo Beitr. z. [Eenntniss der Arten der G'omutalen," Wiegm. Archiv,
1843, i. p. 132.

2 Dissertatio sistens species cognitas Astcriarum. Lundas, 1805.

3 The specific name " peetinata" Linnasus, included both Retzius's specimens and the Decacnemus of Linck.
These belong, however, to two very different types of the genus Comatula, and must be carefully distinguished from
one another. The former is, as above mentioned, an Actinometra, while the latter was called Ahcto by Jfuller,
being simply the common Comatula mediterranea, Lam. Pennant, Adams, and others naturally employed the
Linnean name for this last species ; but Dujardin, following Eetzius and Muller, applied it also to Retzius's original
specimen, which is really an Actinometra, and not an Antedon, like C. mediterranea. This has given rise to much
confusion in the synonymy of these two species.

4 Wiegiminn's Archiv, 1843, i. p. 133. ° ' Gattung Comatula,' pp. 12, 13.
6 Ibid. p. 9.

12 ME. P. H. CARPENTER ON THE GENUS ACTINOMETEA.

weichung allerdings vorhanden, cr liegt ganz zur Seite, cloch ist dies niclit die Ursac h
des Unterschiedes, es gibt vielmehr auch Comatulen von der gewblmlichen Anordnung
der Purcben, Lei denen gleichwohl der Mund seitlich, die Afterrohre central steht. Pig.
von C. multiradiata (PL I. fig. 4). Die fragliche Abweichung beruhrt vielmebr darauf,
dass die fiinf Purcben nicbt symnietriscb fur die fiinf Gruppen der Arme vertbeilt
werden, sondern dass von den fiinf Purcben einzelne herrschend werden und Aeste an
die meisten Arme abgeben. Indein diese Hauptfurcben, nachdem sie die Scbeibe
umzogen, sicb wieder annahern, so entstebt der Scbein eines Cirkels. An in Weingeist
aufbewahrten Exemplaren siebt man indcss, dass es kein gescblossener Cirkel ist;" and
further on (p. 10) be says, " Icb werde daber bei den Arten wo fiinf centripetale Purchen
beobachtet sind, den NamenJfec/o in Klammer clem Gattungsnamen Comatula beifiigen,
wo aber weniger Purcbcnstiimme den excentrischen Mund erreichen, den Namen Acti-
nometra demselben Gattungsnamen Comatula folgen lassen. Also z. B. Comatula
(Alecto) europcca; Comatula (Actinometra) Solaris"

Miiller does not appear to bave been acquainted with De Prcminville's name of An-
teclon ; but be distinctly states tliat Leacb's genus Alecto was constituted a year earlier
than Lamarck's Comatula. He bad in bis earlier communications employed the two
indifferently and as equivalent to one anotber ; but wben be became acquainted with tbe
type represented by Comatula Solaris and elevated it into a new genus, or rather sub-
genus, Actinometra, in contradistinction to Alecto, be retained Lamarck's name Coma-
tula, probably on account of its being so well known, and employed it to designate the
genus in which he included the subgenera Alecto, Actinometra, and Comaster.

Thus the sole cbaracter by which Midler distinguished the first two of tbese subgenera
from one anotber was tbe number of groove-trunks reacbing tbe peristome, irrespective of
the position of tbe moutb. It is tberefore easy to understand that, as many of the spe-
cimens which he examined were dry, and as in otbers, although preserved in spirits, tbe
arms were contracted over tbe disk so as completely to conceal it, be was unable satis-
factorily to determine more than three species of Actinometra. Two of these, Act. Solaris
and Act. Wahlberghii, have been already mentioned; tbe tbird was tbe small Comatula
rotalaria of Lamarck.

(§9) The Aster ias pectinata of Retzius, which presented the same " Bildung des
Schcitels ' as Act. Solaris, resembled it so greatly in other respects, that Miiller
regarded the two as almost identical, or, at any rate, as presenting only varietal
differences 1.

lie seems also to have come to the conclusion that the other species described by
Betzius, the Asterias multiradiata, Linn., had a prior claim to this specific name over
either of the similarly named types described by Goldfuss (Comaster) or Lamarck (Co-
matula multiradiata) ; for be described it as " Comatula (Alecto) multiradiata nob."3.
It is difficult to understand why he called it Alecto; for he had already3 described this

1 'Gattung Comatula,' p. 52. - Ibid. p. 25.

3 Wiegniann's Arehiv, 18-13, i. p. 133, and ' Gattung Comatula,'' pp. 9, 10.

ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA. 13

Lund specimen as belonging to the type of his new genus Actinometra, in which not five,
as in Alecto, hut " weniger Furchenstamme den excentrischen Mund erreichen."

Midler further examined a specimen of the " so-called " C. multiradiata in the Bonn
Museum ; and although he did not actually include it in his type under that name, yet
he seems to have been inclined to do so ; for he says1 that " es stimmt durch den Besitz
der Syzygien an den Axillaria der Anne mit Comatula multiradiata Betz., " hut adds,
" Maul excentrisch, 5 Burchen der Scheibe sammeln die Burchen der respectiven
Arme und kommenam Mund zusammen." (See Bl. I. fig. 4.) Here, again, it is evident
that Midler's description of Alecto will not hold good ; for according to his own descrip-
tions, the Lund and Bonn specimens of Comatula multiradiata, Mull., however much
alike in other respects, differ so greatly in the distribution of the ambulacra on the disk
that one is Actinometra and the other Alecto.

Miiller also referred three specimens contained in'the Baris collection to this type ; and
he was perhaps thinking of the condition of the ambulacra presented by them when he
added the following sentence to his previous description of the Lund specimen, and
named the type Alecto 2. " Mund excentrisch, aber an Weingeistexemplaren ergibt sich
dass die fiinf zum Munde fiibrenden Burchen sich ganz symmetrisch fur die fiinf Gruppen
der Arme vertheilen." This arrangement, which he called the "gewohnliche Anordnung
der Burchen," had been already3 figured by him as occurring in C. 'multiradiata, which,
as he says, differs from the ordinary C. mediterranea in the excentric position of the
mouth (PL I. figs. 1, 1).

It is thus evident that, according to Midler's own nomenclature, two types, differing
only in the "Bilduug des Scheitels," but almost precisely similar in every other respect,
viz. the Lund specimen, on the one hand, and the Baris specimens, on the other, were
referred by him to the same species, Alecto multiradiata, Miiller. It will, however,
be shown further on that the distinction drawn by Miiller between Alecto and Actino-
metra is not a real one, and that the Lund and one of the Baris specimens, both of which
have an excentric mouth and a central or subcentral anal tube, really belong to one and
the same species, Actinometra multiradiata.

(§ 10) Bor a short time after the publication of Midler's Comatula-memoirs the genera
Alecto and Actinometra remained as he left them, both being regarded as subordinate
types of Lamarck's genus Comatula.

A singularly minute fossil species, discovered by Bhilippi 4 between the valves of an
Isocardia cor from the Sicilian Tertiaries, was named by him Alecto alticeps because of
the height of its " Kelchstiick," a character found both in Alecto Eschrichtii and in
A. phalangium, as Midler had already pointed out. A few new fossil species of a more
or less doubtful nature have been since described, and variously referred cither to Mid-
ler's family Comatulinae or to new and distinct types.

The typical genus of this family, Comatula, Lam., has undergone numerous changes in
its definition. Bcemer, who at first revived Linck's name Decacuemus, subsequently

1 'Gattung Comatula,' p. 29. : Ibid. p. 26. 3 Ibid. p. 9.

4 " Alecto alticeps, n. sp., eine tertiiire Comatula-Art von Palermo," Neues Jahrb. fur Mineral. 1S44, p. 540.

14 MR. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA.

withdrew it in favour of Comattda1, the" Knopf," or centrodorsal piece of which was
described by him as an " Ueberrest der verkiimmerten Saule," while, like his prede-
cessors, he mentioned the absence of basals. The existence, however, of external basals,
both in Solanocrmus and in Comaster, led him to regard them, like Midler, as generically
identical ; and he used the name Comaster for this type in preference to Solanocrinus,
as it " bezieht sich nicht nur auf einen lebend und vollstandig bekannten Typus, sondern
druckt auch die Veiwandtschaft richtig, wie Solanocrinus unrichtig, aus." About the
same time D'Orbigny 2, and, a few years later Pictet 3, transferred the name Comatula to
this last-mentioned type, in which the basals appear externally ; while they revived
Linck's name Decacncmus (or, as they named it, Decameros) for the Antedon of De Fre-
minville and the Alecto of Leach. They characterized the genus as only differing from
Comatula, in their sense, in the total absence of the five basals, so that the radials
rest directly upon the centrodorsal piece. Fortunately, however, this peculiar inversion
of the nomenclature employed by Midler was not destined to last ; for in Bronn's
' Klassen und Ordnungen des Thierreichs '4, all the above genera are united into one,
Comatula, which with Glenotremites and another doubtful fossil constitute the family
Comatulkke. Saccocoma and MarsupUes are restored to the places originally assigned to
them by Midler, in special groups, Costata and Tessellata respectively, among the
unstalked Crinoids ; Avkile Eugeniacrinus, which Goldfnss regarded as nearly related to
Solanocrinus, is placed with a few similar forms in a family Eugeniacrinidae, which, toge-
ther with the Pentacrinidae and Apiocrinida?, make up the group Articulata of Midler.

(§ 11) The family Cornatulidse was considerably enlarged a few years later by Du-
jardin and Hup6s, who included in it, as D'Orbigny and Pictet had already done, not only
the tribes Comatuliens and Saccocomiens, but also the Eugeniacriniens, which both the
above authors had ranked among the stalked Crinoids, while MarsupUes, which they
referred to the Cornatulidse on account of its calyx being free, was transferred to the
Cyathocrinidae by Dnjardin, who could " ne voir qu'un caractere secondaire dans
l'absence d'une tige chez plusieurs de ces Crinoides." He distinguishes the three tribes
as follows : — "Nos trois tribus seront suffisamment caracterisees : la premiere, cede des
Eugeniacriniens, par son calice adherent ou pedoncule, jamais libre ; les deux autres,
dont le calice est libre a l'etat adulte, se distinguent parce que celle des Comatubens
porte cirrhes ou rayons dorsaux, dont la derniere, celle des Saccosomiens, est censee
depourvue."

The position of Eugeniacrinus does not concern us at present. Let us now investigate
the species included by Dujardin in the tribe Comatnliens. Under this head he ranks
three genera, viz. Comatula, Lam., Actinometra, Midi., and Comaster, Ag., using the
latter name in the sense in which it was employed by Midler and Rcenier, namely as
equivalent to Solanocrinus. Dnjardin's genus Comatula, however, is not precisely
equivalent to that of Midler, who included in it the two genera or subgenera Alecto and

1 Lethaea geognostica, iii,c Auflage, 1851, Theil iv. p. 133, and Thcil v. p. 177.

- Cours element, de Paleontol. et de Geol. stratigraph. 1S50-1852, vol. '2, i. pp. 138, 139.

s Traite de Paleontol, (Paris, 1857) vol. iv. p. 288. 4 Band ii. Aktinozoen (1SG0), p. 233.

5 Hist. Nat, des Zoophytes, Echinodermes, (Paris, 1SG2) p. 186.

ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA. 15

Aclinometra. The latter was erected into a separate genus by Dujardin, who limited the
application of the name Comatula to those forms only which had been described as
Alecto by Midler — those, namely, in which five main groove-trunks reach the mouth,
irrespective of its position, to which Midler seems to have attached no importance as a
character of any systematic value ; so that Dujardin, following him, says of the mouth
of Comatula (i. e. Alecto) that it is only " ordinairement au centre " \

Further, Dujardin, though really employing the name Actinometra in the same sense
as Midler did, does not describe it in the same way ; he takes no account of the number
of groove-trunks reaching the peristome, to which Midler attached so much importance,
but simply says2, " Ce genre ne differe guere des vraies Comatules que par la position de
l'anus au centre et de la bouche au bord du disque. II en resulte que les gouttieres
ambulacraires, au lieu de se rendre a la bouche en suivant la direction des bras comme
chez les Comatules, s'inflcchissent et suivent le contour du disque." Dujardin adds,
with perfect truth, that the distinctive characters of Actinometra are hardly yet suf-
ficiently established. It will be shown, further on, that his definition of the genus is
really the correct one, and that we must refer to it all those forms of Alecto (Comatula,
Dujardin) in which, as described by Midler3, the anal tube occupies the middle of the
disk, " so dass der Mund seitlich gegen den Rand der Scheibe riickt, ohne dass die
Ambulacra ihre symmetrische Vertheilung auf die 5 Armstamme einbiissen " (PL I. fig. 4).
As Midler had only employed the names Alecto and Actinometra to designate sub-
ordinate types of the Comatula of Lamarck, it is rather unfortunate that Dujardin
should have erected the latter into a separate genus, in contradistinction to Comatula,
and restricted this name to the Alecto of Midler ; for we now know, as mentioned in
section 5, that most of the species described as Comatula by Lamarck belong really to
Actinometra, not only in the somewhat limited sense in which this name was used by
Midler, but also in its wider application as employed in this memoir.

Thus, for example, Midler stated expressly i that Lamarck's original specimen of
C. Solaris in the Paris Museum is identical with the large Vienna specimen, also bearing
the name of C. Solaris, Lam., which he made the type of his new genus Actinometra.

Dujardin, however, paid no attention to this identification of Midler's, and described
the two specimens as C. Solaris and Act. hmperialis respectively, simply on the basis of
Mailer's original diagnosis, published before his visit to Paris. Dujardin thus made
not only two different species, but also two different genera, out of the same type, while
he makes a third species out of the Asterias pectiuafa of the Lund Museum, which
Midler regarded as a variety of Actinometra Solaris.

I have examined a considerable number of specimens of this type, and find it to
exhibit an enormous range of variation in minor points, such as the number and relative
proportions of the cirrhus-segments, the colouring, the presence or absence of a faint
keel on the dorsal side of the arms, &c, and am convinced that none of these can be
regarded as of specific value. A number of such varieties group themselves around a

1 Op. tit. p. 104. 2 Op. tit. p. 208.

3 " Nachtrag zu der Abhandl. iiber die Comatulen," Monatsb. der Berlin. Akad. 1S4G, p. 177.

4 'Gattuug Comatula,' p. 13.

16 ME. P. H. CARPENTER ON THE GENUS ACTINOMETEA.

type possessing certain definite characters, by which it may he distinguished from other
types forming the centres of similar groups of varieties ; but the characters above men-
tioned are usually so excessively variable within each group, that it becomes utterly
impossible to make any use of them as specific distinctions, as Dujardin has done.

Dujardin seems to have detected Muller's oversight in classing the Asterias multi-
radiata of Retzius as an Alecto after previously describing it as an Actlnometra ; for he
transferred it to this genus under the name of Actlnometra multiracUata, and adopted
Muller's specific designation multifield for the original specimens described as Comatula
multiracUata by Lamarck. The third form to which this name has been applied, viz. the
C. multiracUata of Goldfuss, was regarded by Dujardin as a separate genus on account
of its possessing external basals, or, as he called them, " interradials ;" and he restored to
it the old name of Comaster, which had been given up by Midler, including in this genus,
as Roerner had previously done, all the species of Solanocrinus.

(§ 12) Muller had, as we have seen above, referred hoth Alecto and Actlnometra to the
one genus Comatula, while Dujardin limited the application of the latter name to the
species of Alecto only, and gave up the name of Alecto altogether, as had been previously
done by D'Orbigny and Pictet. This was a step in the right direction ; for, as Muller
had already pointed out, this name had been employed since 1821 to designate a section
of the Polyzoa established by Lamouroux. It is a pity, however, that Dujardin, instead
of limiting the application of Lamarck's name Comatula to the species of Midler's sub-
genus A lecto, did not revert to the old name of Anteclon, which was proposed by his
countryman De Ereminville in 1811, and had since received but little notice. This step
was taken by Mr. Norman l a few years later. He did not, however, use Anteclon as
precisely equivalent to Alecto, but applied the name to those forms only in which the
mouth is central and the anus lateral ; and he has been followed by nearly all the sub-
sequent writers upon the Crinoids.

The etymology of Anteclon is somewhat obscure. De Preminville described his typical
species as Anteclon gorgonia, which gives no information as to the gender of the name.
Mr. Norman, however, arrived at the conclusion that it is masculine, and hence
described the common British species as Anteclon rosaceus. In this respect all the later
writers have agreed with him with the exception of Pourtales 2, who employs Anteclon
as a feminine name; and in this step he has since been justified by the result of the
recent etymological researches of Mr. Spedding 3.

It will be used in the same manner in the following pages, both because this seems to
be ctymologically correct, and for the sake of convenience ; since, as long as Muller's
system of trinomial nomenclature is employed for the Comatula, it is far simpler to
write Comatula (Anteclon) rosacea than Anteclon ?°osaceus= Comatula rosacea. In any
case, we are now acquainted with so many different types, Anteclon, Actlnometra, Co-
master, Phanogenia, and 02)hiocrinus, to all of which Lamarck's designation Comatula

1 " On the Genera and Species of the British Echinodermata," Ann. and Mag. Nat. Hist. ser. 3, vol. xv. p. 98.

2 Bull, of the Mus. of Comp. Zool. vol. i. No. 6, " Contributions to the Fauna of the Gulf Stream at Great Depths,"
p. 11 1 ; and Xo. 1 1, " List of the Crinoids obtained on the Coasts of Florida and Cuba in 1867, 18G8, 18GSJ," p. 355.

J ' Nature,' vol. xv. p. 366.

ME. P. II. CARPENTER ON THE GENUS ACTINOMETRA. 17

is equally applicable, that this last can only he used to designate the family, while one
of the two names, Antedon and Alecto, which have precedence over it in point of time,
has gradually become more limited in its meaning, and the other has ceased altogether
to be applied to the Crinoids.

II. On the Characters of the Genus Actinometra.

(§ 13) We have seen that while the distinction drawn by Midler between Alecto and
Actinometra depended upon the number of groove-trunks reaching the peristome, irre-
spective of the position of the mouth, the genus Antedon, as defined by Mr. Norman,
and as subsequently used, is distinguished by having the mouth central and the anus
lateral.

There are, however, numerous species of Alecto in which, according to Miiller1, " die
Afterrohre nimmt die Mitte der Scheibe ein, so dass der Mund seitlich sesen den Rand
der Scheibe riickt ohne dass die Ambulacra ihre symmetrische Vertheilung auf die fiinf
Avmstamme einbussen." These forms have obviously no place in the genus Antedon,
while they were excluded from Actinometra by Miiller, who goes on to say, " In
andcrn abweichenden Arten geht die gleiche Vertheilung verloren, indem der excentriseh
liegende Mund weniger als fiinf Furchen der ambulacra aufnimmt, dann werden
einzelne dieser Furchen herrschend unci verasteln sich, indem sie einen grossen oder den
grossten Theil der Scheibe umziehen, auf mehreren Armstammen zugleich, so dass die
Scheibe von einem Furchenkreis umgeben ist, der jedoch an einer Stclle nicht gcschlossen
ist (Actinometra) (PI. I. figs. 2, 3, 5).

In PL I. figs. 6-16 is represented the distribution of the groove-trunks or ambulacra
on the disks of the eleven specimens of Act. polymorpha which I have been able to
examine. A glance at these, no two of which arc alike, will suffice to show that within
the limits of one and the same species there may occur individuals, some of which would
have been referred by Miiller to Alecto, some to Actinometra, and some which, strictly
speaking, have no place in either of these genera.

Thus, for example, the specimen represented in fig. 16 would probably have been
classed as Alecto by Miiller ; but although five groove-trunks leave the peristome, their
branches are by no means equally and symmetrically distributed to the different arms.
On the contrary, one of them gives off far more branches than any of the others, sup-
plying all the arms borne by two radii (I^-Eo), together with half of those of another
radius (AJ ; while another trunk running straight, out from the peristome bifurcates but
once, and only supplies two of the arms of one radius (Bo). Again, in fig. 15, only four
groove-trunks leave the peristome, one of which gives off a large number of branches, as
in fig. 16, also supplying all the arms of two radii (D^Ea). According to Midler's system,
therefore, this individual is an Actinometra .

In all the other nine specimens of this species which I have examined, however, there
are invariably more than five groove-trunks running out from the excentric peristome
(PL I. figs. 6-11). Even in the small specimen with thirteen arms, represented in fig. 6,
there are six groove-trunks, while in fig. 11 there are eight, and in all the other figures

1 Berlin Monatsberichte, 18iG, p. 177.

SECOND SERIES. — ZOOLOGY, VOL, II. 3

18 MB. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

either six or seven. In no case are there only five with their branches so regularly
distributed as Midler figured them in Alecto multiradiata (PL I. fig. 4) ; nor in the single
individual with only four primary trunks (fig. 15) is the distribution so regular and
symmetrical as in Midler's figure of Actmometra WahlbergMi (fig. 3). Farther, the dis-
tribution of the ambulacra on the disk of the specimen of Act. Solaris, represented in
PI. I. fig. 5, is by no means so symmetrical as Miiller found it to be in the large Vienna
specimen wbich he made the type of his new genus Actmometra (fig. 2). It can hardly
be said of fig. 5 that the " Furchen der zehn Arms miinden in gleichen Abstiinden in
eine die Scbeibe umziehende Cirkelt'urche."

The above instances, which could be multiplied indefinitely, suffice to show the
impossibility of classifying the Comatulce according to the distribution of the ambulacra
on the disk. We have already seen (sect. 9) that Miiller found the Lund and Paris
specimens of his species Com. multiradiata to agree in every respect but this ; so that, had
he adhered strictly to his own system of classification, he would have had to refer the
former to Actinometra and the latter to Alecto. In this case, however, as in all the
specimens represented (PL I. figs. 2-16), there is one point of agreement, viz. the relative
positions of the mouth and anal tube. In the Paris, Bonn (fig. 4), and Lund specimens
of C. multiradiata, Mull., in both the specimens of Act. Solaris, represented in PL I.
figs. 2, 5, in A. Walhberghii (fig. 3), and, lastly, in all the eleven specimens of A.polymorpha
(figs. 0-16), the centre of the disk is occupied by the anal tube, and the mouth is situated
excentrically, either close to the margin of the disk (fig. 11), or at some point rather
nearer to the centre.

(§ 14) After arriving at the conclusion that in this character, the central or excentric
position of the mouth, lies the real distinction between Antedon and Actinometra, and
that the number of groove-trunks reaching the peristome is a character of very minor
importance, I wrote to Dr. Liitken, of the University Museum, Copenhagen, upon the
subject, and was not surprised to learn that he had held this opinion for some time past.
With his usual kindness he has permitted me to make use of the following extract from
an unpublished MS. of his, containing descriptions of some new species of recent
Comatulce : —

" One of the reasons why it is so difficult to identify Midler's species is, that he does
not always mention the positions of the mouth and anal tube, and the direction of the
ambulacra on the disk, but has evidently established a somewhat unnatural distinction
between the differences which may occur in these characters. Two cases may occur : in
the one the mouth is subcentral (' quite central ' probably never occurs), and the ambu-
lacral farrows convennni? from the arms unite into five trunks, which all run directlv
towards the mouth along the shortest line ; they differ, therefore, but slightly in length ;
and the ' interpalmar ' areas defined by them are of almost ecmal size, that containing
the anal tube having sometimes, however, a slight preponderance in size, especially when
the anal tube is placed close to the mouth, almost centrally. In the other case the
mouth is removed towards the margin of the disk ; and of the ambulacra, those only which
come from the arms nearest to the mouth run directly towards that orifice, while the
others, and especially the two enclosing the anal area, are obliged to make a large

MB. P. H. CAEPEXTEE ON THE GENUS ACTLXOATETEA. 19

deviation, and reach the mouth, after a circuitous course, parallel to the margin of the
disk. It is often difficult to state the number of amhulacral furrows abutting ou the
mouth, as they frequently unite immediately before reaching it ; in different species, and
in different specimens of the same species, I have counted 4, 5, 6, 7, 10 stems ori-
ginating from the mouth ; this difference, therefore, is of no importance at all. The
anal tube in all these species [Actinometra, mihi] has a central or subcentral position,
and the anal area occupies the larger portion of the disk. Using this difference as
* fundamentum divisionis,' I have never encountered a doubt whether any type should
be referred to Antedon or to Actinometra, although I have examined a great number of
specimens and species. Moreover the lower or oral pinnules of Actinometra arc always
very different from the others, being flagelliforrn and presenting a more or less distinct
serrature or comb (pinnuloe oralcs prehensiles) ; while in Antedon they are only slightly
differentiated from the others, or are transformed into strong rigid spines, forming a
protective covering over the disk \_A.protectus, mihi]. It will, perhaps, be thought
improper to elevate these sections into genera, as the fossil Antedons would usually not
be generically determinable 1 ; but they are at least very good subgenera for the distri-
bution of the numerous species. The mode of classification here proposed is concerned
with the main point of that established by Midler, but is evidently an amelioration of
it. In Actinometra he describes the five ambulacra as partially uniting before reaching
the mouth, so that their number becomes reduced to three or four [C. Solaris, PI. I.
figs. 2, 5, and A. Wahlberghii, PI. I. fig. 3], while he refers to Alecto all those specimens
in which five ambulacra separately reach the mouth, even though this orifice be quite
excentric and marginal, and the length of the ambulacra therefore esceedin^lv different,
as in Alecto multiradiata [PI. I. fig. 4]. This mode of distinction used by Midler is,
however, very unnatural, and often quite arbitrary or illusory. It is the marginal or
subcentral position of the mouth that is of importance; and this character is never
ambiguous. C. multiradiata [PL I. fig. 4] is not less a true Actinometra than A. soh
[PL I. figs. 2, 5] and A. Wahlberghii [PL I. fig. 3]."

It will be seen from the above note that Dr. Liitken considers Antedon and Actino-
metra as two subgenera of Comatula, Antedon having a subcentral mouth and but slightly
differentiated oral pinnules, while in Actinometra the mouth is excentric and the oral
pinnules bear a terminal comb (PL III. figs. 1-3). At the time I received Dr. Liitken's
note I had had no opportunity of examining any large collection of Comatula?; and his
statement that the oral pinnules of all Actinometra were marked by a terminal comb
was therefore new to me. I have since been able to examine a considerable number of
Comatula?, and, like Dr. Liitken, have never had the least difficulty in determining to
which type any given specimen shoidd be referred, while at the same time I have always -
found that in Actinometra, or Comatula? with an excentric mouth, the oral pinnules bear

1 It -will be shewn further on that there are very decided differences in the shape of the calyx in the two genera
Aatedon and Actinometra. These render the determination of fossil Comatulai (and also of recent specimens from
which the disk is lost) less impracticable than Dr. Liitken supposes.

2 The above passage was written early in 1877. Since then I have examined the large collection of I
brought home by the Challenger.' Out of nearly fifty species with an eccentric mouth, all but two have a terminal
comb on the oral pinnules.

3*

20 ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETBA.

a terminal comb. These two characters, however, do not always coexist; forPonrtales1
describes Antedon meridionalis, A. Ag., as having an excentric mouth, while he says
nothing about the first pinnule, except that it is "rather long, the first five or six joints
webbed by the perisome."

Again, in many of the Comatula with an excentric mouth which I have examined,
the terminal comb is not limited to the oral pinnules only, but may occur at intervals on
different pinnules till near the end of the arms, although it is never so well developed as
it is on their basal or oral pinnules, fewer of the terminal segments bearing the processes
which go to make up the comb.

Loven2 has found the same to be the case in the new Comatula which he has described
under the name of Phanogenia typica. Speaking of the pinnules, he says, on p. 232 : —
" In nonnullis (omnibus ?) articuli 8 1. 9 ultimi convoluti, pectiuati, margine cujusvis ex-
terno in laminam lanceolatam niagnam erectam producto ; " but he goes on to say (p. 233).
" Os centrale. Tubus analis crassus in media area interradiali. Sulci tentaculiferi fere
quales in Antedone."

Here, therefore, a terminal comb on the pinnules coexists with a central mouth ; so
that all the four possible variations may occur of these two characters, viz. the position of
the mouth aud the condition of the terminal segments of the pinnules.

Thus Antedon rosacea Sac. have a central mouth and no comb.
„ Phanogenia tijpica lias a central „ and comb.

„ Actinonielra Solaris kcAmvean excentric „ and comb.
„ Comatula meridionalis lias an excentric ,, and no comb.

Leaving Phanogenia out of consideration for the present, as it was unknown to
Johannes Midler, the following scheme will represent the relations of Antedon, and Act i-
nometra as used by Dr. Liitkeu and myself, to Alecto and Actmometra as used by
Midler :—

Alecto. Ambulacra symmetrically | mouth central. Antedon. Oral pinnules not specially

distributed on the disk. J mouth excentric. "^ distinguished.

Actinomctra. Ambulacra unsymmetri- ] fActinometra. Oral pinnules nearly

cally distributed on the l mouth excentric. | always have a ter-

disk. J minal comb.

(§ 15) "We are now in a position to investigate which species of the numerous Coma-
tulcB described by Midler can be referred to Actinometra under its new definition, and
what further subdivisions of the genus are possible according to the principles of classi-
fication introduced by Midler. Before doing so, however, it will be advisable to devote
a little time to a consideration of the descriptive terminology which he employed, and
of the manner in Avhich it has been modified by later writers.

In Pentaerinus and Comatula Miiller regarded the arms as starting directly from the
five radial axillaries. The two primary arms borne by each of these might either remain

1 hoc. cit. No. 11, p. 355. Pourtales bcre evidently uses Antedon as equivalent to Alecto, and not in the sense in
■which it was employed by Mr. Norman, viz. to designate those forms only in -which the mouth is central (or nearly
so) and the anus lateral. To avoid confusion, I shall speak of this species simply as Comatula meridionalis.

■ " Phanogenia, ett hittills okiindt slagte af fria Crinoideer," Ofver. af Kongl. Yetensk.-Akad. Forhandl. 1SGG,
No. 0, p. 223.

ME. P. II. CAEPENTEE ON THE GENUS ACTINOMETEA. 21

simple, as in Ant. rosacea, or divide more or less frequently into secondary, tertiary,
&c. arms, as in Act. multiradiata and in Pentacrinus; and every segment, like the
radial axillary, preceding- a bifurcation, was called by Muller a "bracbial axillary."

In some of the Tessellate Crinoids, however, the arms do not become free at the radial
axillary, but " der Kelch setzt sich noch weiter fort ; die lladien zerfallen dann in zwei
Distichalradien mit radialia distichalia, die jedes mit einem distichale axillare endeu,
wie bei Actinocrinus moniliformis und Eucalyptocrinus " \ In this case the distichal radii
represent the primary arms of Comatula and Pentacrinus, though Muller never used
the name " distichals " in his descriptions of the species of Comatula; for, as in the
Tessellata the segments composing two adjacent distichal radii are united laterally with
one another by intermediate plates, he regarded them as forming a part of the calyx,
and considered the arms of this group as starting from the distichal axillary, and not
from the radial axillary, as in the Articulate Crinoids.

The two primary arms, or distichal radii, borne upon a single radial axillary, were
called by Muller a " distichium ; " and the interval between two successive distichia
dorsally between the calcareous segments, or ventrally between the corresponding
grooves on the disk, was spoken of by him as " mfcrpalmar,'5 while the interval between
the two primary arms or distichal radii borne by the same radial axillary, or, as Muller
called it, "die Kluft eines Distichiums," was " interbrachial " or " intrapalmar."

The words " interambulacral," " interradial," and " intertentacular," have been also
used by Muller and others to designate the interpalmar areas on the disk of Comatula.
Either of these is preferable to " interpalmar," for reasons which will presently appear,
though " intertentacular " is not universally applicable, as in certain Actinometrce the
posterior ambulacral grooves bounding the large area in which the anal tube is situated
are not provided with tentacles at their sides.

The term "interbrachial" is decidedly preferable to "intrapalmar," which was used
by Muller to designate the small areas on the disk, bounded by the two branches of
each of the five primary groove-trunks. " Intrapalmar " does not convey any clear idea
of the relation of these areas to the divisions of the skeleton, while "interbrachial" dis-
tinctly indicates that they correspond to the intervals between the two primary arms
borne by every axillary radial.

Reenter2 adopted Muller' s nomenclature for the fossil Crinoids, and, like him, con-
sidered that the distichal radii, when present and united laterally to one another, formed
a part of the calyx ; so that the arms were regarded by him as commencing from the
axillary distichals, while he distinguished their different divisions simply as rami of the
first, second, and third order. De Koninck and Le Hon 3, however, regarded the arms
as commencing from the first bifurcation, i. e. from the axillary radial, whether they
become free at once or whether they remain united with the calyx for a longer or shorter
distance. Nevertheless they distinguished the arm-segments by different names in the
two cases, using the expression " pieces brachiales " for the distichals of Muller, i. e. for
those segments which are immovably united with the calyx, while they gave the name
" articles brachiaux " to the movable segments, the brachials of Muller.

1 Baudes Pentacrinus, p. 31. " Lethoea geognostica, Band i. Theil 2, pp. 210, 215.

3 Recherches sur les Crinoides du Terrain Carbonifere de la Belgique, (Bruxelles, 1854) p. 69.

22 ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA.

This view, although unquestionably correct in the case of the Articulate Crinoids
( Comatula, Pentacrinus, &c.), is, as Roenier has pointed out, beset with some difficulties
in its application to the fossil Tessellata ; and Schultze x accordingly reverted to the
original view of Miiller, saying, " Die Arme (brachia) beginuen unveranderlich da, wo
eine dcutliche Gelenkfacette eines festen Kelckstiickes, ihren TJrsprung anzeigt." In
describing the divisions of the arms, he speaks of the brachial axillaries of the first,
second, and third order, without giving them any special names. These are perhaps
scarcely necessary when the number of segments between each division varies so much
in different specimens and in different arms of the same specimen as it does in many
fossil Crinoids, and in Pentacrhtus. Among the Comatulce, however, the number and
character of the segments between the successive divisions of the arms exhibit variations
which are to a certain extent constant in different species, and thus give us the means of
classifying them into larger or smaller groups.

Miiller has availed himself of this character to a certain extent in the scheme which
he gives3 of a classification of the Comatulce according to the presence or absence of
syzygia in the various brachial axillaries ; but though, in his descriptions of the different
species, he furnishes the material for carrying this classification much further, and for
separating species which, in his scheme, stand very near to one another, he never made
anv use of it, simply classifying the Comatulce in the groups which he had constituted,
according to the number of their arms — 10, 20, 40, or more. Under these circumstances
he woidd have been puzzled where to place Act. polymorphs, in which I have found the
number of arms to vary from 13 to 39.

(§ 16) It has been already stated that the arms proper of Comatula begin from the
radial axillaries. In many cases they are united by perisome as far as their second or
third division; and in Act. multifida this perisome contains numerous small calcareous
plates, which render the union of the arms with one another and with the calyx some-
what firmer than usual ; but they are never so united as to be immovable, as their
various segments are connected with one another, except, of course, at the syzygia, by
muscles and ligaments. There is one point about the nature of this union which has
not, I think, received sufficient attention ; and as it shows clearly that the arms of
Comatula and Pentacrinus begin from the radial axillaries, it is worth considering here.
It is this : the first and second segments beyond every axillary, whether radial or
brachial, are nearly always united together in the same manner as the second and third
(axillary) radials.

Thus, for example, in Act. Solaris, and in the forms allied to it, the second and third
radials are united by a syzygy. The same is the case with the first and second brachials.
In Ant. rosacea, and in the various species which are closely allied to it, there are no
muscles between the second and third radials ; but their opposed articular faces
present a vertical and not a transverse ridge, and are so united by ligament that the
two segments are only capable of a lateral movement upon one another, and cannot take
part in any movements of flexion or extension, in which they act as a single segment
only. The first and second brachials are united in precisely the same manner.

1 Monographie der Eclrinodermen des ELflerkalks, (Wien, lSU(i) p. 5.

2 Gattung Comatula, p. 11.

ME. P. H. CAEPENTEE ON THE GENES ACTINOMETEA. 23

In both these groups the primary arms do not subdivide ; so that the total number of
arms is limited to ten ; and we are as yet unacquainted with any Comatula in which the
second and third radials are united by a syzygy and there are more tban ten arms1. This
is, however, the case in Pentacrinus Millleri, in which, in like manner, the first and
second segments beyond every brachial axillary are also united by a syzygy 2. On the other
hand, Pentacrinus asteria, L. (=P. caput-Medusce, Midler), is remarkable for having
muscles between the second and third radials as well as between the first and second 3.
In the same manner the first and second segments beyond every axillary are united
by muscles, and the syzygium is between the second and third segments '.

In nearly all the Comatulce with which we are acquainted, with the exception of Act.
Solaris and the species most nearly allied to it, the second and third radials are united
by ligament only, as in Ant. rosacea, their opposed faces being marked by a vertical
articular ridge (PL VII. figs. 2 b, 3 a, 5 b, 6 a, i). In almost all of these species which have
more than ten arms, the first and second segments beyond every axillary are united by
ligament only, just like the second and third radials 5. Thus in Ant. Savignii every third
segment, so long as the division lasts, is an axillary, and the first and second segments
beyond each axillary are united by ligament only. But in Ant.palmata only two segments
follow each bifurcation, the second of which is again axillary ; it is nevertheless united
to the first one by ligaments only. I have found these same two conditions to occur
together in Act. polymorpha (PI. II. fig. 8), in which the normal number of segments
between every two points of division is three (PI. II. figs. 7, 9, 10), of which the third
is axillary Avith a syzygiuin, as in Ant. Savignii, while the second is united to the first by
ligament only. In exceptional instances, however (PL II. figs. 8, 11), the second segment
may be axillary, and united to the first by hgaments only, as in Ant. palmata.

In every case, after the division has ceased, the union of the first and second brachials

1 There are three Comatula in the ' Challenger ' collection which answer to this description. In two of them the
first and second distichals and the first and second brachials are united by syzygies, like the second and third radials.
But in the third species there is a curious exception to the rule. The rays may divide eight times : and in the primary
divisions there are three distichal joints, the first two of which are united by ligaments and not by syzygy. But in
all the subsequent divisions the first two joints beyond each axillary form a syz3'gy, like the second and third radials.

" This agreement between the mode of union of the second and third radials, and of the first and second brachials
respectively, is seen also in Encrinus moniliformis, in which these segments are united by syzygia as in P. MiiUeri.
See ' Fetref. Germ.' Taf. liv. figs. F, G. The same is the case in Rliizocrinus (Sars, ' Crinoides vivants," pp. 1-3, 22).

3 P>au des Pentaermus, p. 30, and Taf. ii. fig. 8.

4 This does not appear, however, to be always the case ; for iliiller described the syzygium as uniting the first and
second arm-segments in the specimen examined by him, while I have found the same to be tho case in a specimen of
this species contained in the Zoological Museum of the University of Wiirzburg, in which there is certainly no
syzygial union between the second and third radials. In Pentacrinus the opposed faces of two elements, which are
united by a syzygium, are simple, and not radially striated as in Antedon. Sars has found this to be the case in
Bhizoorinus also ; but in its predecessor, Apioerinus obeonicus, Goldfuss, the radial striation of the syzygial surfaces
is very distinct (Petref. Germ. Taf. lvii. fig. 5).

5 The ' Challenger ' collection includes two very abnormal species which present a singular exception to the rule.
The rays divide three times ; and the first two segments (distichals) of each of the ten primary arms are united by
ligaments only, like the second and third radials. So far the rule holds good ; but with the next arm-division there
is a new point of departure. The third or axillary distichal bears the secondary arms, which consist of one axillary
segment only. But this segment is itself primitively double, i. c. it consists of two parts united by a syzygy ; and
the first joints of each of the ultimate arms borne by tis axillary a gree with it in being syzygial segments.

24 ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA.

of the free and undivided secondary or tertiary arms is of precisely the same nature as
the union of the first and second segments of the primary arms home hy the radial
axillary. In fact, it is not at all uncommon for one of the primary arms to remain
simple and the other to divide, as in PL II. fig. 9, c, d, e, which shows that the arms
taken in the strict sense of the word, cannot he regarded as commencing from any point
but the axillary radials.

(§ 17) In practice, however, it is more convenient to regard the arms of Comatula as
beginning from the last bifurcation, i. e. from that axillary the two branches borne by
which do not further divide, but remain composed of a series of simple brachial seg-
ments (bl} b2), &c. In the ten-armed Coma tit lee the brachials are, of course, borne directly
by the radial axillary. But in those forms, such as Act. mvltiradiata, in which the sub-
division of the ten primary arms is carried to a very great extent, it is most convenient
to regard as brachials only the segments of the ultimate branches borne by the last
axillaries, and to give special names to the segments composing the primary and
secondary arms ; for we have already seen that the number of the segments composing
these arms, i. c. between every two successive axillaries, varies in different species, and
it consequently becomes desirable to have some system of nomenclature by which
these differences can be briefly indicated. Under these circumstances, therefore, the
term distichals may be applied to the segments composing the ramified primary arms of

e Articulate as well as of the Tessellate 1 Crinoids, but onlv on the distinct under-
standing that they are really arm-segments and do not enter into the formation of the
calyx, as in the Tessellata ; so that the name is purely a conventional one, employed for
greater convenience in the description of species.

Supposing the secondary arms borne by the distichal axillaries to divide again, we
may consider them as composed of two, three, or four palmar 2 segments, of which the last
is a " palmar axillary " (figs. 10, 11, p.a.), and bears two tertiary arms. These may either
remain simple and composed of brachial segments, or they may continue to divide more
or less frequently. The latter case, however, is somewhat rare ; for if complete series of
distichals and palmars be developed on each radius, the total number of arms rises from
10 to 40 ; and there are not many Comatula? in which this number is exceeded.

If we apply this nomenclature to the species of which mention has already been made,
we should describe Antedon Savignii with 20 arms, as characterized by the presence of
three distichals composing each primary arm and bearing the brachials directly, while
in Antedon palmata, with 30 or 40 arms, there are only two distichals, which are fol-
lowed by two palmars in the secondary arms. Act. polymorpha, again, normally has
three distichals and three palmars (PI. II. figs. 7, 10), while Act. multlrad'iata, Mull., has

1 /. e. As understood by MM. do Koninck and Le Hon.

- I have been accustomed for some years past to use the term " palmar " to designato the secondary arms of the
Crinoids. Professor Huxley, in whose lectures I first heard it used in this sense, informs me that he believes it to
have been so employed by Miillcr ; but I have searched in vain through Midler's works for any definition of the term.

In his description of tho Tessellate Crinoids, however, he describes the plates which continue the series of inter-
radials and interaxillaries in a peripheral direction as " interpalmaria;" and as these partially correspond to the
intervals between the secondary arms, when such arc developed, the latter may perhaps be not incorrectly regarded as
composed of " palmar" segments. It will now be apparent why " interpalmar " is not a very suitable designation
for the strictly " interradial " areas on the disk of Comatula, as was remarked in section 15.

ME. P. H. CABPENTEE ON THE GENUS ACTINOMETEA. 25

only two palmars in each secondary arm, although the number of distichal segments in
the primary arms is usually three. Another proof, if proof were wanted, that the arms
proper of the articulate Criuoids begin from the axillary radial is seen in the fact that
whenever there are three segments in a distichal or palmar series, the second of these,
which is united to the first by ligament only, always bears a pinnule, while the third, or
axillary, is a double segment, i. e. it consists of two primitive segments united by a
syzygium. This is in precise accordance with what we find in all the ten-armed
Comatulce, even in those forms in which, as in. Act. Solaris, the first and second brachials
are united by a syzygium like the second and third radials. In these the second brachial
or the epizygial element of the syzygium bears a pinnule, while the next segment is also
a double one, and corresponds with the compound third brachial of Ant. rosacea and of
the ordinary ten-armed Comatulce. In these last the second brachial is laterally movable
upon the first, and bears a pinnule as in Act. Solaris, while the third has a syzygium (/. e.
is a double segment). This is exactly what Ave find to be the case in those primary and
secondary arms of the multiradiate Comatulce which consist of more than two segments.

(§ 18) The principal character of the genus Actinometra is, as we have seen in sect. 14,
that the mouth is situated excentrically, while the centre of the disk is occupied by the
anal tube. The position of the mouth relatively to the radii or ambulacra, however, is
not the same in all Actinometrce ; thus in Act. Solaris (PL I. figs. 2, 5) the mouth lies
in a radial or ambulacra! plaue, while in Act. WahlbergMi and many other species (PI. I.
figs. 3, 4, 6-16) it is interradial or interambulacral. If we place the disk of an ordinary
Antedon in such a position that the interradial area containing the anal tube is nearest
to us (PI. I. fig. 1), the odd ambulacrum lies in front of the mouth. Let us designate
this as ambulacrum or radius A, and the two branches of its groove-trunk corresponding
to the two primary arms as AL and A2 respectively, A,, being that on the left of the
mouth. Proceeding round the disk in the direction of the hands of a watch, we may call
the other four ambulacra B, C, D, E respectively, and their primary divisions L\ B..
. . . . Ex E2. The anal area is then bounded by the two postero-lateral ambulacra C, D ;
and a plane passing through the mouth and anus, so as to divide the disk into two
symmetrical halves, passes along the odd ambulacrum or radius A, in front of the mouth,
which may therefore be regarded as radial in position.

In Act. Solaris the same is the case, as may be seen in Midler's somewhat diagrammatic
figure (PL I. fig. 2)1, and still better in PL I. fig. 5, which was drawn from a spirit-
specimen, and not from a dry one like Midler's figure. Here, as in Antedon, the odd
ambulacrum is in front of the mouth, which, although excentric in position, lies in the
radial half of a plane which passes through the mouth and anus, so as to divide the disk
into two symmetrical halves. The same is the case in a new Actinometra from the

1 It is not usual to meet with specimens of Actinometra in which the branches of the ambulacral grooves are
distributed with such symmetry as is represented in Muller's diagrams (PL I. figs. 2-4) and in PI. II. fig. 1. Thus,
for example, Midler's figure of Act. Solaris (PL 1. fig. 2) is remarkably regular, much more so than that repre: en1 !
in fig. 5; and I have examined other specimens with more than 20 arms and a radial mouth, in which the regularity
is by no means so distinct as in PI. II. fig. 1. A great range of variation in this respect is seen in PL I. figs. 6-16.
which represent the disks of eleven different individuals of Act. jpolymorpha, no two of which are alike ; the position
of the mouth, however, is constant in all individuals of the same species.

SECOND SERIES. — ZOOLOGY, VOL. II. 4

26 ME, P. H. CAEPENTEE ON THE GENUS ACTINOMETEA.

Philippines (PI. II. fig. 1) with 23 arms. In both these species the dividing plane passes
in front of the mouth between the two primary divisions, A1 A2 of the odd anterior
ambulacrum A, while behind the mouth it is interradial, and separates the two postero-
lateral ambulacra C, B*.

In Act. WahlberghU (PI. I. fig. 3), Act. multiradiata (fig. 4), and Act. polymorpha
(figs. 6-16) the case is different. If, as in Antedon, we designate the two ambulacra
bounding the anal area as C and D respectively, we find that the latter is the odd
ambulacrum, and that a plane cutting the mouth and anus is radial behind the mouth,
in front of which it passes between the two ambulacra A and B ; so that if the centre of
the disk be regarded as the centre of radiation, the mouth lies in an interradial or inter-
ambulacral plane. This is clearly seen when we turn to the dorsal side of the disk, in
which the radii converge to a central point, and not to an excentric one, like the ambu-
lacra of the ventral side. Thus in PI. II. figs. 9-11, the position of the mouth relatively
to the radii is indicated by a X , which in each case is between the two anterior radii
A and B, or interradial.

So far as my experience goes, this type of Actinometra, in which the mouth is inter-
radial and the odd ambulacrum lies behind it, is slightly more common than the simpler
type, in which the mouth is radial and the odd ambulacrum anterior, as in Antedon 1.

(§ 19) We are now in a position to investigate which of the numerous species of
Comatulce described by Miiller belong to the type of Antedon, and which to Actinometra,
and into what groups the latter may be divided according to the principles of classifica-
tion discussed in the last four sections. Midler's specific descriptions afford very little
information in this respect ; for though he says that the mouth, in some instances, is
excentric, and that in others the oral pinnules have a terminal comb, he does not always
do so ; and he makes no use whatever of these two characters in his distinction between
Alecto and Actinometra. In some cases he simply designates a species as Comatula,
without attemping to name it more exactly. This is often, no doubt, simply due to the
fact that, in the specimens which he examined, the arms were so closed over the disk that
he was unable to investigate the distribution of the ambulacra. This, however, is not the
case in the C. trichoptera of the Paris Museum, the disk of which can be readily ex-
amined ; and I believe that Midler did not define this species more precisely because he
was unable to decide whether it should be referred to Alecto or to Actinometra ; for in
one of the two Paris specimens five groove-trunks start from the excentric peristome,
while in the other there are only four. This example alone suffices to show the unsatis-
factory nature of the only distinctive character established by Miiller between his genera
Aleclo and Actinometra, The Paris Museum- contains a very large majority of the

1 Since the above was written, I have examined three large C'omatula-colhctions : — (1) that of the 'Challenger;'
(2) that made by Prof. Semper in the Philippines ; and (3), thanks to the kindness of Dr. (jiinther, that in the
British Museum.

I have, been able to determine the position of the mouth in 80 species of Actinometra. In 4-j of these it is inter-
radial, as in Act. polymorpha : while in the rcmainig 35 it is radial, as in Antedon and in Act. Solaris.

- I would here express my most hearty thanks to Mons. Edmond Perrier, Assistant-Naturalist at the Museum of
Natural History, Jar:! in des Plantes, who lias charge of the Echinoderm collection, and also to his two Assistants, for
the kindness which they showed me during my stay in Taris, and for the readiness with which they afforded me
every possible facility in the prosecution of my work.

ME. P. H. CAEPEXTEE OX THE GENUS ACTIXOMETEA.

27

species of Comatula described by Midler, who personally examined this collection. Last
autumn (1876) I was also able to examine it for myself, and thus to determine which
species should be removed from Midler's genus Alecto and transferred to Actinometra, in
the sense in which this name is understood by Dr. Liitken and myself.

In the following scheme all those species to which no note of interrogation and the
name of no authority is attached, have been determined by myself as Actinometra^, i. e.
as having an excentric mouth and a terminal comb on tbe oral pinnules.

With 10 arms.

Actinometra.

Second and third radial*;
united by a syzygy. <(

Mouth radial.

''Act. Solaris, Miill.
Act. pectinata, Miill.
Act. brachiolata.
Com. purpur a ? 1.

I Com. rosea '.'

y_Act. robusta, Liitken.

Second and third radials f Com. eehinoptera ? -.
united hy ligaments -j
only. Mouth ? (_ Com. meritUonalis'? 3.

With more

than 10

arms.

Second and

third radials

united hy

ligaments

only.

" 2 Distichals.
Axillary has no

syzygy.

<; 3 Distichals.
Axillary has a -^
syzygy.

3 palmars.
Axillary has a V Act. rotatoria, Midi. 4.

syzygy-

/(Act. polymorpha.
Mouth interradial. -< Act. trichovtera.
_ . 1.

2 palmars. ^|
Axillary has
no syzygy.

^ Mouth radial

'2 palmars.

Axillary has

a syzygy.

3 palmars.

axillary has a

syzygy.

[Act. Wahlberghii, Miill.
Act. finihriatn. 5.

Act. multifi&a. <>.

> Mouth interradial . Act. multiradiata.

C Act. parvicirra. 7.
■i Act. polymorpka. S.
[Act. Bennettii, Bohlsche. 9.

Remarks.

1. I have not personally examined either C. purpurea or C. rosea. Muller seemed to
think that the former might be a young condition of Act. Solaris1; so that it is most pro-
bably a true Actinometra. C. rosea, however, presents a difficulty; for Muller says
expressly2 that the first pinnules are uot specially distinguished; but, except in this
point, he regarded it as very closely related to C. brachiolata, which is a true Acti-
nometra.

:Gattimg Comatula? p. 13.

2 Ibid. p. 14.

!:;

28 ME. P. II. CARPENTER ON THE GENUS ACTINOMETRA.

2. O. echinoptera, on the other hand, has, according to Midler1, a very marked comh
on the oral pinnules. He says nothing, however, about the position of the month ; and I
have unfortunately not been able to examine the species for myself.

3. According to Pourtales - the mouth is excentric in Com. meridionalis ; but he makes
no mention of a comb on the oral pinnules. If it should be absent in this species, and
also iu Com. rosea, while it is present in Phanogenia, in which the mouth is central, it
becomes obvious that the only external character, besides the shape of the calyx, on which
we can rely with any certainty in the determination of the generic position of any
Comatula, is the nearly central or the excentric position of the mouth.

4. According to Midler ;1 there are only two radials in Act. rotalaria which are united
by a syzygium, while they bear the distichal axillaries directly ; and these are also syzy-
gial segments. Although, like Miiller, I examined Lamarck's original specimen of this
species, I cannot confirm the above statement. It is true that only two radials are visible
externally; but this is often the case in Comatula?, with a wide centrodorsal piece; and
I A\as quite unable to satisfy myself that they are united by a syzygium, as Miiller
describes, and as is the case with the second and third radials of Act. Solaris, Avhile I was
equally unable to determine a syzygial union between the two segments of which the
distichal series is composed. Lamarck's original specimen of this species in the Paris
Museum is wrongly labelled C. brevicirra, Troschel.

5. The dry specimen of Act. fimbriate in the Paris Museum, from the voyage of
Peron and Lesueur, is labelled C. multiradiata, Lam. ; while Peynaud's original speci-
mens from the Strait of Sonde are labelled C. brevicirra, Trosch., together with Act.
rotalaria and the Vavas variety of Miiller' s C. parvicirra.

6. In Jet. multijida the tertiary arms borne by the palmar axillaries may divide again
several times. In every case there are only two segments between each division, and all
the successive axillaries, like the palmar axillaries, have no syzygia. Although Miiller
mentions this in his description of the species, it is placed in his scheme in a group in
which the axillaries of the arms have syzygia. In reality, however, this is the case with
the distichal axillaries only.

7. The type specimen of this last species does not exist in the Paris Museum under
that name, nor even under C. brevicirra, Troschel, which seems to have been used as
an equivalent for it ; but I believe that three small spirit-specimens from the voyage of
Peron and Lesueur in 1803, which arc classed, with two specimens of Act. pectinata,
under the name of C. simplex, are really those which were described by Miiller as
C. parvicirra.

8. It will be noticed that Act. polymorpha has already appeared higher in the list as
a species in which palmars are not developed. In some individuals but few of the 10
primary arms bear axillaries ; so that the total number of arms is less than 20 (PI. II.
fig. 9) ; while in others all the primary arms divide again, and so do the resulting secon-
dary arms, so that the total number is little short of 40 (PI. II. figs. 10, 11). This is
very unusual ; for I only know of two other species which present the same kind of

1 '(juttuug Comatula,' p. 14. " Loc. cit. No. 11, p. 355.

3 ' Gattung Comatula, ' p. 20.

MR. P. H. CARPENTER OX THE GENUS ACTINOMETRA. 29

irregularity. As a general rule all the individuals of a species agree in the presence or
absence of distichals and palmars.

9. In Act. Bennettii there are more than 70 arms; but all the axillaries are like the
first one (distichal), and not different from it, as in Act. multifield. According to
Miiller1, every fourth segment is an axillary without a syzygium ; but Bohlsche2 has found
this to be incorrect. There are, indeed, four segments between every two points of division ;
but the last two are united by a syzygium ; so that the formula becomes three distichals,
palmars, &c, of which the axillary has a syzygium. Bohlsche's figure of the disk of his
specimen is noteworthy; for though five groove-trunks leave the excentric peristome, as in
Alecto, yet their distribution to the arms is not by any means symmetrical, so that he
seems to have decided upon calling it Actinometra. Midler named it simply Comatula.

(§ 20) In the above scheme are included all the species of Comatula which have been
determined by myself or by others3, as far as I know, to belong to the type Actiuometra.
Fourteen of these were known to Midler ; and of the remaining 23 species described by
him I have been able to refer 16 to Antedon, viz. : —

Ant. adeonte. Ant. macrocnema. Ant. phalangium.

Ant. articulata. Ant. Milberti. Ant. Philiberti.

Ant. carinata. Ant. Milleri. Ant. Reynaudii.

Ant . Eschrichtii. Ant. palmata. Ant. rosacea.

Ant. Jacquinoti. Ant. petasus. Ant. Sarsii.

Ant. Sarignii.

To these must be added

Ant. armata, Pourt. Ant. cubensis, Pcmrt. Ant. Hagenii, Pourt.

Ant. bicolor, Mus. Paris. Ant. dividua, Mus. Paris. Ant. polyactinis, Mus. Paris.

Ant. celtica, Barrett. Ant. Dubenii, Bolsche. Ant. rubiginosa, Pourt.

The following list contains the seven remaining species of Comatula described by
Miiller which I have not been able to examine, and of which I know no descriptions from
which it is possible to obtain any information as to the position of the mouth or
the character of the oral pinnules.

C. Cumingii. C. elongata. C. flagetlata.

C.japonica. C. nova Guinea?. C. tessellata,

and C. timorensis. To which must be added C. brevipinna, Pourt.

III. External Characters of Act. poltmorpha, and Specific Diagnosis of the Type.
(§ 21) In Act. yolymorpha, as in all Actiuometrce, the mouth (PI. I. figs. 6-16, m)

1 ' Gattuug Comatula' p. 28.

: '• Uebar Actinometra Bennettii und erne neue Comatida-Aib {Antedon Dubenii) " Wiegm. Archiv, l^GG, i. p. 90.

3 Dr. Liitken has named several new species of Actinometra besides Aet. robusta — for example, Act. tenant; and
Act. trachygaster. But his descriptions have not, as far as I know, been published ; and I have had no opportunity of
examining any specimens of his new species except Act. robusta ; so that I am unable to llace them in the classi-
fication given in the previous section.

Grube has described a new Actinometra from Borneo, and two new species of what be calls Comatula. His de-
scriptions (Jahresber. d. Schlesisch. (Jesellsch. 1375, Nat.-Hist. Sect. pp. 54, 55) are, unfortunately for me, not to be
obtained in this country.

30 MR. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA.

does not occupy a central or subcentral position on the ventral surface of the visceral
mass as it does in Antedon, but is placed more or less excentrically, and may be some-
times almost marginal (PL I. fig. 11). It occupies the centre of the peristome, P, and is
bounded by two lips, a large anterior and a smaller posterior one ; so that its opening is
very inconspicuous, and usually so much extended in a direction transverse to the
antero-posterior diameter of the disk, that it presents the appearance of a simple slit, as
is well seen in PL II. fig. 2.

The circumoral portion of the peristome, or the peristome proper, is a more or less oval
depression in the ventral perisome of the disk, which completely surrounds the oral
opening, and gives origin to the ambulacral grooves or, more shortly, the ambulacra.
Beneath this depression lies the water- vascular ring which gives off a trunk imder
each of the ambulacra radiating from it. The number and distribution of these are
very variable, as is seen in PL I. figs. 0-16. This principally depends upon the
way in which the ambulacra divide, so as to give rise to the groove-trunks corresponding
to the ten primary arms. As a general rule, the two ambulacra corresponding to the
radii T) and E unite into one large posterior trunk, from which the branches are dis-
tributed to the various arms into which these radii divide (PL I. figs. 8-10, 12-10). In
other cases the left lateral ambulacrum, E, leaves the peristome alone (figs. 6, 11) ; while
in others it is partially united with the posterior ambulacrum. D, its anterior division,
E2, leaving the peristome by a separate trunk, while its posterior division, E„ unites with
the posterior ambulacrum (fig. 7).

As a general rule, the right lateral ambulacrum, C, leaves the peristome alone, and
supplies the arms of the corresponding radius ; but in figs. 9 & 15 it is seen to unite with
the posterior division, B2, of the right anterior ambulacrum, B.

The mode of division of the two anterior ambulacra is excessively variable. As a
general rule there are no principal trunks corresponding to the two radii A and B, and
the primary divisions, A13 A,, B^ B,, start directly from the peristome. In the specimens
with but few arms, however, each pair may be united for a longer or shorter distance
(PL I. figs. 0, 7), as in Antedon (fig. 1). Not unfrequently the posterior divisions
Al5 B2, of these two anterior ambulacra unite for a longer or shorter distance with the
two large ahoral groove-trunks, to form an open horseshoe-shaped curve bounding the
anal area (figs. 12, 15, 10). The position of the anal tube in this ai'ea, and also with
regard to the whole surface of the disk, varies somewhat with the position of the mouth ;
it is rarely, if ever, absolutely central. Its appearance differs very much according
as it is full or empty : sometimes its aperture is so completely closed as to be scarcely
discernible, though the tube below is widely distended ; and sometimes the aperture is
patent with its edges everted and crenate, and the tube leading to it quite shrunk and
flaccid (PL II. fig. 2).

(§ 22) In Antedon the median line of the ventral perisome of all the arms is occupied
by an ambulacral groove, with a floor of ciliated epithelium. This groove extends also
on to all the pinnules, wdth the exception of those borne by the second distichals and
second palmars, &c. (when present), and by the lowest brachial segments. Beneath it
lie the radial water-vascular and blood-vascular trunks, between which last and the

ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA. 31

ciliated epithelium of the floor of the groove lies a fibrillar structure, to which I have
given the name of the " subepithelial band " l, and to which a nervous character has
been attributed by myself and by all the other observers who have described it. Each
side of the ambulacral groove is bounded by an elevated fold of perisome, the edge of
which is not straight, but cut out into a series of minute valvules, the cresceutic or
respiratory leaves (Wyv. Thomson), or " Saumlappchen " of the German authors.

At the base of each leaf, and to some extent protected by it, is a group of three
tentacles, one of which, the more distal one, is larger than the other two. This trind
group of tentacles and the cavity of the respiratory leaf adjacent to them receive a
common branch from the radial water-vessel. These groups of tentacles alternate on
the opposite sides of the ambulacral groove from the base to the tip of each arm, and are
distributed in the same manner at the sides of the ambulacra of the disk, though they
are not so markedly developed, especially near the peristome, where every lateral branch
of the water-vessel supplies only one tentacle. The crescentic leaves at the sides of the
groove are also far less distinct than in the arms, the edges of the folds of perisome
bounding the groove being only marked by a faint wavy line, and not distinctly cut out
into " Saumlappchen."

In many Actinometrce, however, the above description only applies to the arms of the
two anterior radii, A, B (PI. II. figs. 2, A), and to more or fewer of the antero-lateral
arms, C^ and E^. The arms of the posterior radius, D, and of the posterior divisions of
the lateral radii, C2 and El5 are often entirety devoid of tentacles ; and in many of them
the ventral perisome not only exhibits no ambulacral groove, but is, on the contrary,
convex, as in the oral pinnules of Antedon (PL II. figs. 5, 6).

We have just seen that in Act. pohjmorplia, as in all Actinometrce with an interradial
mouth, the anal area is bounded by two large aboral groove-trunks, which start from the
posterior angles of the peristome, and form a horseshoe-shaped curve, tbe limbs of which
are unequal in size (PL II. fig. 2). The smaller right limb is formed by the right lateral
ambulacrum, C ; while the larger left limb represents the posterior ambulacrum, D, com-
bined with part or the whole of the left lateral ambulacrum, E. In neither of these limbs
are the tentacular groups and crescentic leaves so well developed as they are in the
two anterior ambulacra. After the branches to the two antero-lateral primary arms,
C^ and E2, have been given off, or sometimes even sooner (PL I. figs. 13, 15), the ten-
tacles at the sides of the two aboral groove-trunks become more and more insignificant,
and finally disappear altogether, while the position of the crescentic leaves is only
indicated by a very faint wavy line at the edge of each groove.

In small specimens with but few arms (PL I. figs. 6, 9) the grooves of the posterior
(D) and postero-lateral arms (C2, E,) may remain in this condition; but in larger speci-
mens with many arms all trace of the crescentic leaves disappears, and the two edges of
the groove meet and unite so as to produce the condition represented in PL II. figs. 5 & 6,
where the ventral surface of the arms and pinnules is convex, and does not show the
least trace of a groove of any description.

" Kcmarks on the Anatomy of the Arms of the Crinoids. Part I.,'' Journ. of Anat. and Physiol, vol. x. p. 57!'.

32 ME. P. H. CARPENTER ON THE GENUS ACTINOMETEA.

The position of the point at which the two folds of perisonie hounding the sides of the
original ainbulacral groove meet and unite, varies extremely. The fusion may, though
rarely, take place on the disk ; sometimes it is at the hase of the arms, and sometimes
not till near their middle or terminal portions. In any case, however, the fusion, when-
ever it occurs, is so complete that all trace of the original ainbulacral groove is entirely
obliterated.

(§ 23) The bearings of this fact upon the different views advanced by Greeff l and
Ludwig 2 respecting the homologies of the ainbulacral grooves of the Crinoids will he
best discussed at a later period, when the changes undergone by the various structures
underlying the grooves are described and illustrated. One point, however, must he
noticed here on account of its importance with respect to the two views now entertained
regarding the nervous system of Comatula.

As long ago as 1865 it was stated by Dr. Carpenter 3 that the cord which traverses the
length of the arms between the subtentacular and cceliac canals, " and which was regarded
by Professor Miillcr as a nerve, really belongs to the reproductive apparatus. But it
will also be shown that a regular system of branching fibres proceeding from the solid
cord (described by Professor Miillcr as a vessel) that traverses the axial canal of each
calcareous segment of the rays and arms, is traceable on the extremities of the mus-
cular bundles ; and reasons will be given for regarding these fibres as probably having
the function of nerves, though not exhibiting their characteristic structure." During
his residence in the Philippine Islands, Professor Semper had also discovered that the
arm-nerve of Miillcr is really a part of the generative system ; and in a short paper *
published some time after his return he announced this fact, and suggested at the same
time, " dass der lusher immer als Gefass aufgefasste Strang im innern des Kalkskelettes
ein Nervenstrang sei, und dann ware wohl das im Kelch liegende sogenannte Herz als
ein Ganglion anzusehen."

These observations of Dr. Carpenter's and Professor Semper's were unfortunately
overlooked for many years, so that even as late as 1871 Midler's erroneous statements
with regard to the nervous system of Comatula were repeated in the valuable text-book
of Gegenbaur 5 and in many smaller works. At the commencement of 1876, however,
two very different views respecting the nervous system were put forward nearly simul-
taneously by Greeff and by Dr. Carpenter. The former 6 described the whole floor of the
ambulacral grooves on the arms and disk of Ant. rosacea as constituting a radial
nervous system, starting from an oral nervous ring in the peristome, and corresponding

" Peber den Bail der Echinodermen. III. Mittheilung," Sitzungsb. der Gesell. z. Beforder. d. gesamm. Naturwiss.
zu Marburg, 1S72, No. 11, p. 155.

- - Beitr. z. Anat. der Crinoideen," Nachrichten von der Konigl. Gesells. der "Wissens. u. der G. A. Universitat
zu Giifctingen, 1876, No. 5, pp. 107, 10S.

3 " Researches on the Structure, Physiology, and Development of Antedon rosaceus. Tart I.," Philos. Trans,
vol. clvi. p. 705.

J " Kurze anatom. Bemerk. iiber Comatula," Arbeit, aus d. zool.-zootom. Inst. zuWurzburg, Band i. (lS74),p. 262.

5 Grundriss der vergleich. Anat. p. 222.

6 '• Peber den Bau der Crinoideen," Marburg. Sitzungsb. 1»70, No. 1, Jan. ^^6, p. 21.

MR. P. H. CARPENTER ON THE GENUS ACTINOMETRA. 33

in position and histological structure with the typical Echinoderm nerves. At the same
time he denied the nervous nature of both the structures described as nerves by Miiller
and Dr. Carpenter respectively, viz. the genital cord, the so-called " rachis," on the one
hand, and the axial cords in the centre of the calcareous segments on the other.

A week after the publication of Green's views, Dr. Carpenter 1 announced his belief
that a complicated apparatus, " consisting of the outer cylinder of the Crinoidal stem, of
the five-chambered central organ formed by the dilatation of that axis within the centro-
dorsal basin, and of the cords proceeding from it to the arms and eirrhi," should be
regarded as the central portion of a nervous system. This view was based both upon
anatomical and upon physiological considerations : —

(a) That while a siugle arm may be made to coil up by irritating one of its pinnules,
the whole circlet of arms closes together when an irritation is applied to the pinnules,
which arch over the mouth — an act which affords a strong indication of the "inter-
nuncial" action of a definite nervous system.

(b) That stimulation of the central quinquelocular organ ("heart" of Miiller and
Grceff) contained in the calyx, with which the axial cords of the arms are in connexion,
is followed by sudden and simultaneous flexion of all the arms.

(c) That these axial cords give off successive pairs of branches, which ramify upon the
muscles connecting the arm-segments.

Shortly after the announcement of these views on the part of Dr. Carpenter, Ludwig 2
described a ventral nervous system as existing in Comatula in common with all the
other Echinoderms. He attributed a nervous character, not to the whole epithelial floor
of the ambulacral grooves, as was done by Greeff, with whose researches he was unac-
quainted, but to a fibrillar layer beneath it, and more or less distinctly separated from
it. This layer, which was also discovered independently by myself3 and Teuscher4,
and was regarded by us both as of a nervous nature, is the " subepithelial band " men-
tioned in sect. 22. Ludwig, like Greeff before him, denied the nervous character of the
dorsal axial cords of the arms ; Teuscher discussed it as possible, but hesitated to accept
it on account of the morphological difficulties involved in such a view.

Baudelot5, who seems to have been unacquainted Avith Dr. Carpenter's earlier state-
ments, was apparently struck with the nature of these cords, though he could not regard
them as nervous. After describing; their structure and their union in the calvx to form
the pentagonal commissure, he adds, "Ainsi done chez les Comatules il existe des parties
qui eciclemment rC appartiennent point au systeme nerveux (!), et qui dans leur disposition
aussibien que dans leur structure offrent une analogie presque complete avec les cordons
nerveux des autres Echinodermes." It must be remembered that Baudelot wrote before
the discovery of the so-called " ventral nerve " of Comatula ; but, in any case, I do not
quite see the force of his " evidemment."

1 " On the Structure, Physiology, and Development of Antedon rosacew" Troc. Roy. Soc. no. 160, Jan. 20th,
1876, pp. 219-226.

- Gottingen Nachriehten, no. 5, Feb. 23rd, 1S70, p. 106.

3 Journ. Anat. Phys. x. p. 578.

4 " Beitr. z. Anat. der Echinodermen, I. Comatula mediterrawa," Jenais. Zeitseh. B. x. p. 253.

b " Contribution a l'histoire du systeme nerveux des Echinodermes," Arch, de Zool. Exp. et Gen. i. p. 211.
SECOND SERIES. — ZOOLOGY, VOL. II. 5

3 ± ME. P. II. CABPEXTEE ON THE GENUS ACTIXOMETEA.

Ill the centre of every segment of the skeleton of Act. polymorpha and of all the other
Comatulce which I have examined, from the first radials to the ends of the arms and
pinnules, and also in the cirrhus-segments, these axial cords increase considerahly in
size, and give off four principal branches. Two of these run towards the ventral side, and
in the calyx disappear in the neighbourhood of the muscles connecting the segments,
though I must confess that I have never been able to trace them any farther (PL VIII.
fig. 3, »'). In the arm-segments, however, they continue their course towards the ventral
surface and break up into numerous branches, some of which, as I have already described1,
extend to the tips of the cresceiitic leaflets at the sides of the tentacular furrow. The
two inferior or dorsal trunks run towards the surface of the skeleton ; and while some
of their branches are lost in the plexus of tissue forming its organic basis, others seem to
become connected with epidermic structures in a manner which will be described at
length further on.

Not one of the German observers makes any mention of these branches, although two
of them at least have examined Antedon Eschrichtii, while they have all cut sections of the
arms of species of Actinometra, in which genus I find them to be particularly distinct.
It is obvious that the facts above stated strongly support the view expressed by Dr. Car-
penter and by myself, that the axial cords of the arms are of a nervous nature ; and the
experiments made by Dr. Carpenter 2 at Naples have shown conclusively : —

1. That the quinquelocular organ is the instrument of the perfect coordination of the
swimming movements of the arms, which involve the conjoint contraction of several
hundred pairs of muscles.

2. That nothing contained in the visceral mass is essential to the perfect coordination
of the swimming-movements, and that therefore the subepithelial band or arnbulacral
nerve of the German authors has no immediate relation to those movements, even if it
be a nerve at all.

3. That section of the subepithelial band in an arm does not prevent its playing its
usual part in the regular swimming-movenuiiLS.

4. That destruction of the axial cord of an arm by the application of acid causes the
arm to become rigidly stretched out, while all the others work as usual.

Since the publication of these experiments Greeff seems to admit the nervous nature
of the axial cords, and of the yellowish fibrillar envelope (PI. VIII. figs. 1-3, N) of the
quinquelocular organ from which they proceed. Ludwig3, however, while allowing
their force, cannot admit the existence in the Crinoids of an antiambulacral nervous
system, of which Ave know as yet no homologue in the other Echinoderms, but sees
no difficulty in regarding the quinquelocular organ, its fibrillar envelope, and the
axial cords proceeding from it, as parts of a blood-vascular system, like that of the other
Echinoderms, although he admits (p. 87) that " ihncn vergleichbare Gebilde sind bis
jetzt bei anderen Echinodermen nicht bekannt gewordert." The axial cords of the

1 Journ. . x. p. 584.

• "Su [omenta] on the Structure, Physiol, and Develop, of Aniedm rosacius," Proc. Eoy. Soc.
no. 169, 1876.

3 "Bcitr. zur Anat. der I ." Sep— Abdruck aus der Zeitsch. f. wissensch. Zool. B. sxviii. Heft 3, p. 81.

ME. P. H. CARPENTER OX THE GENUS ACTIXOITETEA. 35

calcareous segments are regarded by him (pp. 80, 80) as " unverkalkt gebliebene Tbeile
der bindegewebigen Gruudlage der Kalkglieder, deren Aufgabe es ist, aus dem Blutge-
fasssysteni, genauer aus den fiinf Kammern die kemniihrende Eliissigkeit aufzunebmen
unci den Arm- und Pinnulagliedern zuzufiibren."

"Without going into the question as to how far the organic basis of the calcareous
skeleton can be regarded as of a connective-tissue nature, I would only remark that it
is difficult to see why the axial cords, which Ludwig supposes to consist of uncalcified
connective-tissue fibres, should give off branches the terminations of which are entirely
outside tbe skeleton, as is the case with those which reach the erescentic leaves at the
sides of the tentacular groove, and which therefore cannot take any part in the nutrition
of the tissue forming the organic basis of the skeleton.

(§ 24) This is not the place for a full discussion of Ludwig's views on the nervous
system of Comatula ; but one point must be briefly referred to. I have already ' stated
that in some arms, and in most of the pinnules, of many Actinometrce, the subepithelial
band or nerve of Ludwig is entirely absent, and also that "if the axial cords are not
nerves, and if the subepithelial bands are to be regarded as the only nervous structures
in the whole Crinoid organization, the difficulty presents itself that the oral pinnules of
the European Crinoids, and more than half the arms, with the majority of the pinnules
of some forms of Actinometra, are entirely devoid of a nervous supply.

"The oral pinnules of Antedon have been shown by Dr. Carpenter2 to be extremely
susceptible of irritation; when they are touched in the living animal, the whole circlet
of arms is suddenly and simultaneously coiled up over the disk, while irritation of one of
the ordinary pinnules is simply followed by flexion of the arm which bears it.

" The structure of tbese oral pinnules, which are borne in Antedon rosacea by the
second brachials, differs very considerably from that of the pinnules borne by the other
brachial segments ; for not only are they sterile, but they have neither tentacular appa-
ratus nor ambulacral groove, their ventral surface being slightly convex instead of being
concave as in tbe ordinary arms and pinnules. Tbis has been mentioned by Teuseher3;
but he has omitted to state that the ordhiary ciliated epithelium of the ambulacral
groove, with its subjacent nervous layer and nerve-vessel, are also absent."' Ludwig
enthely ignores tbis argument, althongh he confirms the above statement concerning the
oral pinnules of Antedon; in the text he is obliged to confess that "Fraglich est mir
geblieben ob die oralen Pinnuhc einen Zweig des radiaren Xerven besitzen oder nicht "
(p. 75) ; while his figure of a section of an oral pinnule (pi. xvii. fig. 55) entirely con-
firms the statement quoted above, to which, however, he makes no reference.

This condition, which is limited in Ant. rosacea to the oral pinnules, sometimes exists
in whole arms and in all the pinnides borne by them in many species of Actinometra.
Even in the arms which come off from the anterior or oral side of the disk the ambulacral
groove does not give off regular branches to the pinnules borne by the third and succes-
sive brachial segments ; but a variable number of these first pinnules (sometimes only

1 " Remarks on the Anatomy of the Arms of the Crinoids. part ii.,"' Journ. of Anat. and Physiol, vol. xi. October,
1876, p. 89.

■ Proc, Roy. Soc. no. 1G6, p. 226. 3 Jenaische Zeitschrift, x.p. 249.

36 ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA.

three or four, sometimes as many as forty) resemble in this respect the oral pinnules,
their ventral surface being convex, and devoid of any ciliated epithelium or subepithelial
baud ; while their water-vessel is simple, without any lateral extensions to respiratory
leaves and tentacles. In these oral arms, however, branches of the ambulacral groove
enter the pinnules sooner or later, so that the terminal ones are always provided with a
distinct tentacular apparatus, while the floor of their median groove is of the usual cha-
racter, consisting of a ciliated epithelium and a subepithelial fibrillar baud.

We have seen in sect. 22 that in many cases the ambulacral grooves going to the
aboral arms become less and less distinct as they get further and further from the peri-
stome, and that their tentacles diminish and finally disappear. At the same time the floor
of the groove becomes very much reduced in extent, its epithelial layer thinner and
thinner, and the subepithelial band almost invisible, until, in those cases in which the
two sides of the groove meet and unite, the ciliated epithelium and subepithelial band
disappear altogether. Consequently, when this union takes place on the disk, whole arms
are entirely devoid of any nervous supply, if we suppose, with Ludwig, that the anti-
ambulacral axial cords are not of a nervous nature, and that the " subepithelial bands "
are the only nervous structures in the arms. In such cases it would naturally be
expected that these arms would be incapable of performing the regular swimming-move-
ments like those in which there is an open ambulacral groove and a subjacent " ambu-
lacral nerve;" but Professor Semper, who has kept Actinometrae in his aquaria for
weeks together, informs me that he never saw the least trace of any irregularity in the
alternating movement of their arms while swimming.

The gradual obliteration of the ambulacral grooves by the approximation and fusion of
the elevated folds of perisome at their sides, which may occur to so great an extent in
Actinometra, is found also at the ends of the arms and pinnules of Aiitedon Eschrichtii.
Ludwig states (p. 75) that their terminal segments have no ambulacral groove or
tentacles ; and he gives a figure of a section through the end of a pinnule (pi. xiii.
fig. 12), the ventral surface of which is convex, while there is no ciliated epithelium
or subepithelial band (ambulacral nerve), although in the text Ludwig makes no
mention of their absence. I have found the gradual obliteration of the groove in
these cases to take place in precisely the same manner as in Actinometra, the only
difference being that the point at which the sides of the groove meet and fuse is much
further from the disk in the one case than in the other.

If we suppose, with Ludwig, that the subepithelial band is the sole structure of a
nervous nature in the whole Criuoid organization, it is difficult to understand the fact,
which Ludwig himself admits (p. 10), that it gives off no branches except those which
go to the tentacles. It is true that in the Ophiuridea the ambulacral nerve does give
off branches which go to the muscles, besides those proceeding to the tentacles, as
described by Lange \ Teuscher -, and Simroth - ; but the researches of the first-mentioned
observer render it very doubtful whether the representative in the Ophiuridea of the

1 " Beitr. z. Auat. und Histiol. d. Astericn und Opkiuren," Morphol. Jahrb. ii. Heft 2, p. 241.

2 " Bjitr. (fee, II. Opliiurida?," Jenais. Zeitsc.h. x. p. 274.

8 " Ar.at. und Scbizogonie der Ophiactis virens, Rars," Zcitsch. f. wissensek. Zool. xxvii. p. 473

ME, P. H. CARPENTER ON THE GENUS ACTINOMETRA. 37

subepithelial baud of Comatula takes any part in the formation of these branches.
Ludwig further admits that he has been quite unable to find any sense-organs at the
ends of the arms or pinnules of Comatula like those which exist in the Asteridea, and,
in discussing the views of Greeff, expresses it as his opinion (p. 78) that " die subepi-
theliale Faserlage, welche durchsetzt wird von fadenformigen Verliingerungen des dariibcr
gelegenen Epithels allein den Nerven darstellt." There can, I think, be little doubt
that this subepithelial band is of the same nature in the Crinoids and Asterids ; and
it is therefore very interesting that the nervous nature of this structure in the Asterids
has recently been disputed by Lange \ who regards as nervous only some cellular
masses separated from the subepithelial band by a lamella of connective tissue, and
projecting into the lumen of the two nerve-canals. He believes these cell masses to swell
into a large ganglionic mass beneath the pigment-spot ; while, in his opiniou, the sub-
epithelial band, together with the ciliated epithelium and the cuticula, constitutes a pro-
tecting integumentary layer. Lange finds a corresponding condition in Ophiura te.vlu-
rata, in which the radial nervous system is better developed than in the Asterids, and
consists of a series of paired ganglionic masses, connected with one another by transverse
and longitudinal commissures. On the ventral side of this ganglionated cord is a longi-
tudinal band, which Lange regards as the homologue of the protecting integumentary
layer formiug the floor of the ambulacral groove of the Asterids, and which, as is uni-
versally admitted, corresponds to the subepithelial band, epithelium, and cuticula of the
ambulacral grooves of the Crinoids.

Langc's views have been partially accepted by Simroth2; but the correctness of them
is altogether denied by Teuscher 3, who regards Lange's nervous cell-masses in the
Asterids simply as the " geschichtetes Epithel " on the wall of the nerve-canals; while
the terminal ganglionic mass under the eye-spot described by Lange is represented by
Teuscher (pi. xix. fig. 22) simply as a " bindegewebiges Polster." Ludwig1, too, speaks
of the nervous cell masses as local thickenings of the epithelium of the nerve-canals, which
are not present in every species. This is naturally a very strong argument against
Lange's views ; but Ludwig omits to apply similar reasoning to his own opinions regarding
the Crinoid nerves. The subepithelial bands {his nerves) are not constant in every
arm of many species of Actimomeira. Still less do Teuscher and Lange agree about the
nervous system of the Ophiurids ; Lange's ganglionic masses are described as artificial by
Teuscher, who, as in the Asterids, regards as the nerve only the fibrillar structure repre-
senting the subepithelial band of Comatula.

The question is still an open one ; and it is therefore of no slight interest to learn that
the supposed ambulacral nerve, or subepithelial fibrillar band, is not always present in
the arms of Comatula, and that even when it exists it is certainly not motor in function,
even if it be a nerve at all5-,

1 Morph. Jalirb. ii. 274. - Zeitseh. f. wiss. Zool. xxvii. pp. 550-500.

3 " Beitr. &c, III. Astcriden," Jen. eitsch. x. p. 513.

* "Beitrage zur Anatoinie der Aetoriden," Zeitschr. fur wiss. Zool. xxx. p. 191.

s It is worth noticing here that the " ambulacral nerve " of Comatula must bo derived either from the niesoblast
or from the hypoblast of the embryo. It is developed immediately beneath the tentacular atrium of the pentacrinoid
larva, which Gotte has shown to be the most anterior portion of the left peritoneal sac. This is lined by hypoblast,

38 MR. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

(§ 25) We have seen in sect. 22 that in certain of the arms of Actinometra the water-
vessels are simple tubes, like the integumentary water-vessels of the Jfolpadhlce, and are
not in connexion with any tentacular appar/atus. "Whether the mouth be radial or
interradial, the non-tentaculiferous arms are invariably the aboral ones ; so that in the
latter case they belong to the trivium (PI. I. figs. 0-15), and in the former to the bivium
(PL I. fig. 5) \ This last, however, is not always the case ; for I have a specimen of
Act. solaris in which an anterior arm (d) of one of the two ambulacra of the bivium is
tentaculiferous, while a posterior arm (EJ in the trivium has no tentacles ; it is never-
theless aboral in position, as maybe seen from PI. I. figs. 2, 5.

In only one individual of Act. polymorpha (PI. I. fig. 15) have I found a non-tenta-
culiferous arm on one of the two anterior radii (A, B.) ; but this was a very remarkable
case. Out of 31 arms, 19 were entirely devoid of a tentacular apparatus ; and in
15 of these the fusion of the two sides of the ambulacral grooves had taken place cither
on the disk or in the basal arm-segments, so that an " ambulacral nerve " was wanting
in nearly half the total number of arms. In the other four non-tentaculiferous arms the
groove remained open for a short distance, and then closed in the manner above
described. Three of these four arms constituted the anterior division (E2) of the left
lateral ambulacrum ; but the fourth was the first arm of the left anterior ambulacrum (Ax),
and was borne upon the same palmar axillary as a well-developed ordinary tentaculiferous
arm. Pieces of the middle portions of these two arms are represented in PI. II. figs. 3 and 5,
and their terminations in figs. 4 and G. With this exception, I have invariably found the
non-tentaculiferous arms on the aboral side of the disk ; their number and distribution,
however, vary extremely, not only in different species but in different individuals of
the same species.

Thus in Act. poly morpha, in Plate II. fig. 8, the former is as low as -^ of the total
number of arms, while in fig. 15 it reaches Jf. Even in two individuals with the same
number of arms it may be different ; thus in figs. 8 and 9 it is -£q and ^ respectively,
and in tigs. 12, 13 it is ^f and |f . The individual represented in fig. 12 was also
remarkable for the fact that one of its aboral arms belonging to the posterior division
of the left lateral ambulacrum (E,) was tentaculiferous, while those on either side of it
were not so.

In all the specimens of the type of Act. polymorpM which I have examined, and in
three of its varieties, of which I have, unfortunately, only single specimens, more or

and appears to be separated from the hypoplastic epithelium lining the water-vascular ring by a remnant of the meso-
blastic tissue which occupied the blastoccel of the Echinopoedium. One or other of these two layers, the hypoblast lining
the atrium, or the mesoblast between it and the epithelium of the water-vascular ring, must give rise to the " ambu-
lacral nerve,'' which cannot be in any way derived from the epiblast. I am inclined to believe that the " nerve " is
most probably of mesoblastic origin, and that the remainder of the mesoblast (in this position) is converted into the
muscular layer of the ventral wall of the water-vascular ring; whilo the blood-vascular ring is a remnant of the
primitive blastoccel. Huxley ('Anatomy of Invertebrate,' p. 559) has suggested a similar origin for the nerve-canals
(perihamral canals, Ludwig) of the Asterids.

1 In all these figures (PI. I. figs. 5-16) the tentaculiferous ambulacra are indicated by dark lines, and the non-
tentaculiferous grooves by fainter lines.

ME P. H. CARPENTER OX THE GENUS ACTINOMETRA. 39

fewer of tlie arms have no tentacular apparatus ; but in the fourth variety (PI. I. fig. 1G)
all the arms are of the usual character, with open ambulacral grooves fringed with
crescentic leaves and groups of tentacles. I have found the same variation to occur also
in Act. Solaris. In this case the number of arms is limited to ten, which may be all ten-
taculiferous, or from one to four of the aboral arms may have no tentacular apparatus.
[Note. February 1879. — No less than 23 out of 48 species of Actinometra brought
home by the ' Challenger ' have more or fewer grooveless arms. I have cut sections of
these arms in two species, and have obtained the same results as with Act. polymorpha
and Act. Solaris. The "ventral nerve" and ambulacral epithelium are conspicuous by
their absence, while the axial cords in the skeleton give off branches freely in the centre
of each arm-joint, as I have already described for other species, both of Antedon and of
Actinometra. Two points are noteworthy. In one species one of the posterior ambu-
lacra stops quite abruptly on the disk, and the two arms to which it would naturally
have gone, with its " nerve," tentacles, &c, receive no branches from any of the adjacent
grooves to supply the deficiency. Again, in one of the largest Comatulce 1 have ever
seen (a ' Challenger ' specimen from the Philippines) there are more than 100 arms,
manv of which are both grooveless and nerveless, as I have found from sections. But
these abnormal arms are not limited to the hinder part of the disk as is usually the case ;
for there are several on each radius.]

The distribution of the non-tentaculiferous arms in Act. polymorpha varies, like their
number, to a very great extent. In any case they always occur on the odd posterior radius,
D (PI. I. fig. 8) ; when more are developed they may occur on the posterior divisons,
G, and El5 of the two lateral radii, C, E ; and they may then be called postero-lateral
(PI. I. figs. G, 12-11) ; and when the proportion of non-tentaculiferous to tentaculiferous
arms hecomes very great, more or fewer of the antero-lateral arms, C13 E2, belong to the
former class (PL I. figs. 7, 9-11, 13), while in exceptional cases non-tentaculiferous arms
may occur even on the anterior radii (fig. 15, A).

(§ 20) The condition of the ambulacral groove and of the tentacular apparatus is not
the only point in which the anterior or oral may differ from the posterior or aboral
arms. The former taper very slowly, contain far more segments, and are much longer
than the latter, while the form of their terminal portions and of the pinnules which
these bear is altogether different (PL I. figs. 4, C). When viewed from the dorsal side
(PL II. fig. 7) the basal portions of the two kinds of arms are precisely similar ; they
widen slightly between the first and second syzygia, *. e. from the third to the tenth
brachial, remaining uniform till the third syzygium on the fourteenth brachial, after
which they hegin to taper. Up to about the twenty-fifth or thirtieth segment the oral and
the aboral arms decrease in width at about the same rate ; but from this point onwards
there is a great difference between them. The arms borne by the two anterior radii,
A and B, taper very slowly, the length of their segments increasing considerably, while
the breadth only diminishes very gradually ; at the same time the middle and terminal
pinnules, in which no genital glands arc developed, become very long and filiform, and
remain so until the last few segments, when their length suddenly diminishes very
considcrahly (PL II. fig. 1).

40 ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

I have never been able to ascertain what is the precise mode of termination of these
anterior arms ; even when the arm ends in such a manner that there is no reason to
suppose that its terminal segments have been broken off, its few last pinnules appear
simply as immature, and the last pair are separated by a delicate prolongation of the
arm-stem, on which no pinnules have been as yet developed. Dr. Carpenter : has found
the same " growing-points " at the ends of the arms of Ant. rosacea, all of which are of
the same character as the oral tentaculiferous arms of Actinometra ; and he was never
able satisfactorily to determine the normal mode of termination of the arms.

With the posterior arms of Actinometra, however, the case is different. From the
twenty-fifth segment onwards they taper very rapidly, and instead of reaching a length
of 145 millims., as the anterior arms with some 150 segments may do, they have only
some 80 segments, and rarely attain a greater length than G0-70 millims.

At the same time their terminal pinnules are little, if at all, longer than those of the
middle portion of the arm (PI. II. figs. 5, 6) ; and the centre of the dorsal half of each
of- their segments is occupied by a dark-brown egg-shaped body, of a peculiar cellular
nature, which I have reasons for believing to be a sense-organ2 (PI. II. fig. 6, o.b).
These bodies commence to appear in the pinnules at about the beginning of the second
third of the length of the posterior arms, and are continued to their extremities. The
pinnules of the last few segments decrease very slowly in size ; and the arm ends in an
axillary segment which bears two pinnules of the ordinary character, each provided with
the brown ovoid bodies or " sense-organs" (PL II. fig. 6, o.b).

These bodies, which may occur, though but rarely, on one or more of the anterior
tentaculiferous arms, do not exist in all the specimens of Act. polymorpha which I have
examined. In three out of my eight specimens of the type they are entirely wanting ;
and they are also absent in all the single specimens of the four varietal forms which
I have investigated. I have also failed to find them in the non-tentaculiferous arms of
Act. Solaris*.

Between these two kinds of arms, the long anterior ones on the radii A, B, with a
wide ambulacral groove and a well-developed respiratory apparatus, and the short
posterior ones of the radius D with a closed groove and no external respiratory appa-
ratus, all possible forms of transition may occur. As a general rule, more or fewer of
the antero-lateral arms, d and E2, are tentaculiferous ; but they never reach such a great
length as the anterior arms, and their terminal pinnules are by no means so long and
slender. At the same time the postero-lateral arms, C3 and El9 although generally non-
tentaculiferous, have, except in rare cases, a more or less open groove for the greater part
of their length, which, while greater than that of the posterior arms of the radius D,
is less than that of the antero-lateral arms of CL and E2 ; and their pinnules increase
slightly in length from the middle till near the end of the arm.

1 Phil. Trans. I860, p. 723, plate xxxviii. fig. 4. = Journ. Anat. & Phys. vols. x. xi. Incc. citt.

3 Sense organs occur in two of the ' Challenger' species — one from Banda (which is probably the young of Act.
jydlymorpha), and one (a new species) from the Admiralty Islands. In both cases they are limited to the hinder arms,
some of which are grooved and others not.

MR. P. H. CARPENTER ON THE GENUS ACTINOMETEA. 41

The arms of Act. polymorpha may thus be roughly classified as follows : —

Anterior, on radii A and B, 120-150 segments. Pinnules increasing in length to the terminal ones,
which are very long and slender. Tentacnliferous.

Anterolateral, on Ct and Es, 100-120 segments. Terminal pinnules long and slender. Tenta-
cnliferous.

Posterolateral, on C2 and E,, 80-100 segments. Terminal pinnules stout, and rather longer than the
median ones. Usually have " sense-organs " and narrow ambnlacral grooves, but are non-tentaculiferous.

Posterior, on radius D, 60-80 segments. Terminal pinnules stout, but shorter than median ones.
Sense-organs. Usually no grooves. Non-tentaculiferous.

Another difference between the anterior and posterior arms is that the genital glands
in the latter are far more developed than in the former. Not only is their number
greater, although the total number of pinnules on a posterior arm may not be much
more than half that of an anterior arm, but they also attain a very much greater size ;
the basal and median pinnules of an anterior arm being very much less swollen than
the corresponding pinnules of a posterior arm. A similar inequality in the development
of the genital glands has been noticed by Alex. Agassiz1 as occurring in the Echini. This
difference in length in the anterior and posterior arms of. Act. polymorphs, and in the
character of their terminal pinnules, seems to be to a certain extent dependent upon
the condition of the respiratory apparatus occupying their ventral surface. When this
is well developed the arm seems to have the power of indefinite growth ; for in the
single specimen (PL I. tig. 16) in which all the thirty-three arms were normal and ten-
taculiferous as in Antedon, there was no very appreciable difference in the lengths of
the anterior and posterior arms-. The shape of the terminal pinnules, however, was of
a slightly different character in the two cases, though the development of the genital
glands was about the same ; and we have just seen that those arms are the shortest in
which the ambulacral groove entirely closes, and the water-vessel is reduced to a simple
tube without any lateral tentacular branches, while it is in these arms only that any
definite mode of termination is known. This may occur before half the number
of segments have been developed which are commonly met with in an anterior ten-
tacnliferous arm.

(§ 27) The ventral surface of some of my specimens of Act. polymorpha is marked by
small calcareous concretions, somewhat resembling the " blumenartige Knotchen mit
mehreren blattartigen Fortsatzen " described by Midler3 in the Vienna specimen of
Act. Solaris. When present, they are usually scattered around the peristome, and

1 ' Revision of the Echini," part iv. pp. 680, 681.

2 Not only are the arms of different, lengths in the ' Challenger ' species of Actinometra, which have ungrooved
hinder arms, but there are three species in which the anterior arms are longest, although all, anterior and posterior
alike, are grooved and bear tentacles. In another species the arms are all grooved and all equal in length, but the
distribution of the syzygia is quite different in the anterior and posterior arms.

:i • Gattung C'omatula,' p. !"_'.

SKCON'U SERIES. — ZOOLOGY, VOL. LI. b"

42 ME. P. H. t'AKPENTEK ON THE GENUS ACTINOMETRA.

disposed along the edges of the primary groove-trunks proceeding from it, and there
are generally some upon the sides of the anal tube. They are particularly well de-
veloped in the dark variety from Ubay, in which all the arms are tentaculiferous.

(§ 28) The "oral pinnules" of Act. polymorpha, those, namely, which arch over the disk
so as to protect it, are borne by the second distichals and second palmars when these are
present, but in any case upon the second brachials, those of the distichals and palmars
being the longest. They are all very long and slender, consisting of some 30 or 40
segments ; and their terminal portions exhibit the peculiar characteristic comb made up
of processes which rise from the outer margin of the ventral surface of each calcareous
segment (PI. III. fig. 2), just as in Act. Solaris and Act. pectmata (PI. III. fig. 1). The
number of segments on which these processes may be developed varies from 10-12 on
a distichal pinnule, to 6-8 on a brachial pinnule ; but in cases in which no second
distichals or palmars are developed, so that the pinnule on the second brachial is the
first of the series, it is much longer than usual, and more of its terminal segments bear
the comb-like processes.

The oral pinnules of the dark Ubay variety of Act. polymorpha differ considerably
from those of the type and of other Actmometrce ; not only are they much stouter, but
their terminal comb is differently constituted (PL III. fig. 3). As is usually the case, the
lower processes gradually develope themselves from the outer margin of the ventral
surface of each calcareous segment ; but towards the end of the pinnule they gradually
come to rise less and less from the outer margin, and more and more from the median
portion of the ventral surface of each segment, until finally, on the last two or three
segments, they are developed from the inner margin. Consequently the comb, when
viewed from above, is seen not to lie altogether on the outer side of the pinnule, as is
usually the case, but to start from the outer side, cross its ventral surface, and finally
come to lie on the inner side of each pinnule, i. e. on the one nearest the arm.

Both in the type of Act. polymorpha and in all the four varieties, the pinnules
diminish in length from that of the second distichal (when present) to those borne by
the fourth and fifth brachials; that of the sixth brachial is longer, and usually contains

well-developed genital gland, so that it is slightly swollen. Prom this point onwards
the pinnules increase in length till about the thirtieth brachial, after which their length
and character vary according as the arm is tentaculiferous or non-tentaculiferous.

(§ 29) The dorsal aspect of Act. polymorpha differs from that of most Antcdons, and
especially from that of Ant. rosacea, in the fact that the plane of the second and third
radials, like that of the first, is parallel to the vertical axis of the calyx, and not inclined
to it, as in Antedon ; so that the dorsal surfaces of the whole of the pieces of the calyx
lie in one horizontal plane. The centrodorsal piece is circular (PI. II. figs. 9, 10, c d), or
pentagonal (fig. 11), and conceals a large portion of the pentagon formed by the first
radials, less in young specimens with but a few arms (fig. 9) than in large and full-grown
specimens with many arms (PI. II. figs. 10, ] I & PI. VI. fig. 2). It is usually a flattened
plate with a slight concavity in the centre of its outer surface; and around its margins

ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA. 43

are disposed some 20 or 25 cirrhi in one row, but with occasional traces of a second,
in which the cirrhi alternate in position with those of the first row. The number of
segments in each cirrhus is normally from 11 to 14, of which the last forms a recurved
claw, while a more or less distinct spine is usually visihle upon the dorsal edge of each of
the three or four penultimate segments (PI. III. figs. 8-11).

In PI. II. fig. 8, is seen an abnormal condition of the centrodorsal piece, which is of
an irregular oval form, and so extended as to conceal large portions even of the second
radials. These last are usually more or less completely united with one another late-
rally. The amount of their union is to a certain extent dependent upon the number
of arms developed. Thus in the small specimen with only 13 arms, represented in
PL II. fig. 9, the second radials are not united laterally for more than half their length;
in fig. 10 (2(3 arms) the union is somewhat more complete, and even more so in fig. 8
(28 arms), while in the variety with 39 arms, represented in fig. 11, the second radials
are completely and closely united with one another all round. This rule, however,
appears to he only a specific one, and not generally applicable to all Comaiulce ; for in
the SO-armed Phanogenia the second radials, as figured by Loven ', do not appear to
be united with one another any more closely than they are in the small 13-armed speci-
men of Act. polymorpha (PI. II. fig. 9).

In Act. polymorpha the two segments (first distichals, palmars, or brachials) borne by
any axillary are united to one another laterally to about very much the same extent as
the second radials are ; i. c. when the number of arms is small, their first segments,
whether primary, secondary, or tertiary, are not laterally united in pairs with such com-
pleteness as when the division of the ten primary arms is carried to any considerable
extent (PI. II. figs. 8-11, dlt pu &J.

When the arm-division is unequal it is generally carried further in the trivium or
posterior radii, C, D, E, than in the two anterior radii, A, B, which form the bivium.
This is well seen in PI. II. fig. 9, in which no distichals are developed on either of the
two anterior radii ; and again in fig. 10, in which, while distichals are developed all
round, the division is carried no further in one of the anterior radii, while in each of the
others from one to three palmar series may be developed. In only four normal cases
have I found an anterior radius to bear more arms than a posterior one. In each of
these the total number of arms was considerable, and one at least of the two posterior
radii bore the same number of arms as the abnormal anterior one. Thus, for example, in
PL II. fig. 11, each of the radii bears eight arms, with the exception of the posterior one
(D), on which only seven are developed. This, however, is an abnormal case of fracture
of the whole radius between its second and third segments. The new portion is con-
siderably smaller than the old, the proximal articular face of the new axillary being far
less wide than the corresponding distal face of the old second radial ; while both the
distichal series which it bears are imperfect and abnormal, so that the absence of a
further division in one of the secondary arms is not. particularly remarkable.

(§ 30) The number of arms that may be developed in Act. polymorpha is a character

1 " Phanogenia," loc. cit. p. 230.

SECOND SERIES. — ZOOLOGY, VOL. II. 7

41

3IB. P. H, CAEPENTEE OX THE GENTTS ACTLXCttlETEA.

of extreme variability. In the specimens I have examined it varies from 13 to 39 ;
so that, with one remarkable exception (PI. II. fig. 8), the ten primary arms do not, at
the most, divide more than twice, while in two specimens with 18 and 13 arms
respectively two and seven of the primary arms remain undivided. I believe, how-
ever, that, as a general rule, an axillary is developed on each primary arm, and that the
amount of further division is variable, but that a tertiary division is probably excep-
tional, so that the number of arms in this species will be found rarely to exceed 40.

It will have been already apparent from the position assigned to Act. polymorpha in
the classification given in sect. 20, that I consider the typical number of distichals and
palmars in this species to be three, of which the second (d,) bears a long pinnule, while
the third or axillary segment (d a) consists of two primitive segments united by a
syzygium. A typical specimen of this condition is seen in PI. II. fig. 10. Out of the
twelve specimens of this species which I have examined, but four others resemble this
one in having all then distichal and palmar series regularly developed. In each of the
other seven specimens one or more of the distichal or palmar series is irregular, con-
sisting onlv of two segments, the second of which is axillary without a svzvgium. In
one very remarkable case, represented in PI. II. fig. 8, one of the palmar series is
reduced to a single segment placed on the distichal axillary, being also itself an
axillary bearing the brachials directly on one of its articular surfaces, while on the other
are two segments which may be called suprapalmar, of which the second [sp.a) is an
axillary without a syzygium, and bears two arms.

Excluding this remarkable case, the comparative frequency of the usual variations in
the distichal and palmar series in the twelve specimens of Act. polymorpha examined by
me is seen in the accompanying Table. Prom this it appears that out of 111 distichal

Table I. — Showing the Variations in the Distichal and Palmar Series.

Specimen.

Total
number

of
Arms.

Bis-iehal Series. Palmar Series.

Total
n umber.

Of three

segments.

Of two
segments.

Total
number.

On two distichals. On three distichals.

Of two
segments.

Of three
segments.

Of~two

segments.

Of tl

segments.

I.

II.

III.

IV.

V.

VI.

VII.

YIII.

Yar. 1
2
3

4

Total

13

1-
20
25
26
28
28
31

20
29
39
33

3

8
10
10
10
10
10
10

10
10
10
10

3

4
5
10
10
7
10
10

10

10

8

9

4
5

'3

"2
1

5
6
5

8
11

ii
19
13

2

1
1

3
1

2
1

2
6
5

7
11

9
14
11

310

111

96

15

70

2

2

1

65

ME. P. H. CAEPEXTEE OX THE GEXTJS ACTIXOATETEA. 45

series 96 were normal, and that out of 70 palmar series 65 were normal, i. e. consisted
of three segments, of which the second bore a long pinnule, while the third was axillary,
with a syzygiiun. The three forms of variation exhibited by the abnormal palmar
series are of considerable interest, because some of them, at least, represent the normal
condition of the palmars in other groups of Actinometrce. Thus the most frequent one,
two palmars on three distichals, is typical for Act. multifida, while that of three palmars
on two distichals is typical for Act. rotalaria. The third variation, two palmars on two
distichals, occurs in Act. tenax, Ltk , and in a few new ' Challenger ' species ; and it is
typical in several species of Antedon — for example, in. Ant. palmata and Ant. articulata.
Specimens Xos. II. and III. are remarkable for the fact that the numbers of regular and
irregular distichal series are in each case equal to one another; so that a specific diagnosis
based upon either of these specimens alone, would, as is evident from the above Table,
have been entirely incorrect.

The amount of variation in these characters is so enormously great that only after
examination of a considerable number of specimens is it possible to draw conclusions
of any value respecting the use which may be made of these characters for systematic
purposes. The above Table, however, will, I think, show clearly that I am justified
in assuming the normal number of both distichal and palmar segments in this species
to be three, of which the second bears a long pinnule, and the third is axillary with a
syzygium.

(§31) The same variability occurs in the position and distribution of the syzygia
in the arms, but, as might be expected from the nature of the case, to an infinitely
greater extent. In most of his specific diagnoses Miiller gives the position of the first
svzv^iuin on the arm and the average number of segments which occur between everv
two successive syzygia throughout the rest of the arm. Only in a very few cases
does he make mention of the position of the second syzygium, which I believe to be
a character of nearly or quite as great systematic value as the position of the first ; and,
owing to its greater constancy, of considerably greater value than the number of seg-
ments between every two successive syzygia, which I will call the " syzygial interval."
It will be seen from Table I. that the total number of arms in the 12 specimens of
Act. polymorpha at my disposal reached 310 : 11 of these were broken below the third
segment ; but of the remaining 299, the first syzygium was on the third brachial in
283 cases ; and in 156 of these the second svzv«ium was on the tenth brachial. The
irregnlarities in the position of the first syzygium were limited to tbree specimens,
and, as will be seen from Table II., nearly all confined to one variety.

46

ME. P. H. CAEPENTEE ON THE GENES ACTINOMETEA.

Table II. — Showing Irregularities in the Position of the first Syzygium.

No.

Irregular series of Syzygia.

Type.

Variety.

III.

VI.

4.

Brachials.

1.

On 4, 9, 13.

. .

1

'2.

4, 10, 14.

2

3.

5, 13, 16.

1

4.

6, 10, 14.

1

5.

9, 13, 17.

2

6.

10, 14, 16.

1

7.

10, 14, 18.

1

4

2

8.

10, 14, 19.

1

Total

2

4

10=16

In nearly every case the irregularity appears to have been the result of regeneration,
the arm having been broken, either in the distichal or in the palmar series, or between
the third brachial and the preceding axillary, and a new one developed with an irregular
syzygial series ; although in many cases similarly regenerated arms of other specimens
exhibit perfectly normal series of syzygia. One of these unusual cases, in which there is
no syzygium on the third brachial (b3) and the first syzygium occurs on the tenth seg-
ment (610), which is usually the position of the second syzygium, is seen in PI. II. fig. 8.

We have seen that when the first syzygium is on the third brachial, the position of the
second is in the great majority of cases on the tenth brachial ; that is to say, the first
syzygial interval is six simple segments, while the second and all the subsequent intervals
are, as a general rule, only three simple segments, though the range of variation on cither
side of this number is very great.

Table III. shows the variations in the positions of the second and third syzygia in all
those 283 arms in which the first syzygium is on the third brachial. From the last
column of this Table it is evident that in Act. polymorplia and its varieties the normal
position of the second syzygium is on the tenth brachial, and that in those cases in
which it does not occupy this position it is much oftener on the eleventh or twelfth
segment than on the eighth and ninth ; i. e. that variation, when it occurs, is in the direc-
tion of increase rather than of decrease in the length of the first interval. This is more
clearly seen in Table IV., which shows the number of segments intervening between the
first and second syzygia in all the above cases.

MR. P. H. CAEPENTEE OX THE GENUS ACTIXOMETEA.

47

Table III. — Showing the Variations in the Positions of the second and third
Syzygia in the Arms of twelve specimens of Act. polymorpha.

No.

Positions of the
first three Syzygia.

Type.

Vnr. 1.

Var. 2.

Var. 3.

Var. 4.

Total.

No. of Variations

in the

first interval.

Brachials.

1.

On 3, 4, 5.

1

1

2.

3, 4, 9.

1

1

3

3.

3, 4, 10.

••

•■

1

1

4.

3, 5, 10.

2

2

2

5.

3, 6, 11.

1

1

1

G.

3, 7, 11.

o

o

2

7 •

3, 8, 11.

1

1

8,
9.

3, 8, 12.
3, 8, 13.

1
1

2

o

5
1

8

10.

3, 8, 14.

1

1

11.

3, 9, 11.

1

1

12.

3, 9, 12.

1

1

23

13.

3, 9, 13.

10

h

3

20

14.

3, 9, 14.

1

1

15.

3, 10, 12.

1

1

16.

3, 10, 13.

3

1

4

s

17.

3, 10, 14.

76

10

23

16

13

138

18.

3, 10, 15.

1

1

150

19.

3, 10, 16.

1

1

1

3

20.

3, 10, 17.

3

1

. .

4

21.

3, 10, 19.

1

1

22.

3, 11, 14.

3

1

4

23.

3, 11, 15.

33

4

1

3

2

43

50

24.

3, 11, 16.

3

••

3

25.

3, 12, 14.

1

1

26.

3, 12, 16.

30

1

1

32

34

27.

3, 12, 18.

1

1

28.

3, 13, 16.

1

1

29.

3, 13, 17.

1

1

2

30.

3, 14, 18.

Total number of 1

variations . . J

1

1

2

2

174

20

28

38

23

283

283

48

ME. P. II. CAEPENTEE ON THE GENUS ACTINOMETEA.

Table IV. — Showing the Variation in the number of Segments in the

first Interval.

No. of
Segments.

0.

1.

2

o.

4.

5.

(5.

7.

8.

9.

10.

Total.

Type ....

Var. 1 . . .

2

3 ...

4 ...

1
2

2

1

2

4

2
o

13

i
3

81
14
25
22
14

39
4

I

Q

O

32
1
1

1
1

1
1

174

20
28
38
23

Total

3

')

1

2

8

23

156

50

34

2

2

283

It is also seen in Table III. that even when the second syzygium is abnormally placed,
it is usually the case that the interval between it and the third is the normal one of three
simple segments, so that scries like 3, 9, 13 ; 3, 11, 15 ; and 3, 12, 16, are very common.
This is well seen in Table V., which shows clearly that the length of the second interval
is normally three segments ; that, like the first, it tends to vary in the direction of excess
rather than of defect, and that the range of variation in both directions is greater in the
varieties than in the type of Act. polymorpha.

Table V. — Showing the Variation in the number of Segments in the

second Interval.

No. of

Segment 3.

0.

1.

2,

3.

4.

g

6.

8.

Total number
of Arms.

Typo

Var. 1 ....
2

3

4 ....

i

o

y
1

'4
1

151
16
25
30
21

8
0

3

1
1
1

3
1

1

174
20
28
38
23

Total

1

3

15

243

10

0

4

1

2S3

After the fourteenth brachial a syzygium usually occurs on every fourth segment ; so
that the number of segments composing the syzygial interval is normally three. It is, how-
ever, very unusual to meet with an arm in which this interval is constant throughout
its whole length and does not vary to a greater or less extent. In only seven arms out
of the whole number which I have examined have I found this to be the case, together
with normal first and second intervals, although twenty-three other arms were regular
from the second syzygium onwards. These thirty arms were distributed among five out

ME. P. H. CAEPENTEE ON THE GENES ACTIX03IETEA.

49

of the eight specimens of the type, while in none of the other three was the syzygial
interval constant throughout the length of any of the arms ; the same was the case with
the four varietal specimens.

Table VI. — Showing the Variations in the Syzygial interval (usually 3 segments)
in the Arms of twelve specimens of Act. poly morjpha.

3.

4.

5.

6.

7.

8.

9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
2-6.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.

Number of Segments in
each interval.

1.

1,2.
1,5.

0.

0, 0, 0.

0, 2.
1.
1,
1,
1,
1,

1, 2,2.
1, 4.

1, 11.
o

2, 0.
2,1.

oi l -> ■)

**) X5 "*-> ~ ? — ■

—) x, ^j — •
2, 1'.

.) .. o

— , — , _.

., .; ., ..

— , — . — , — .

■) ■) -. •) g

O ■> .> o .) ., a

— > -> -. -- -- -. -•

2,4.

2, 4, 2.

2, 4, 5, 1.

2, 5.

2, 0.

4.

4 1 2 4

4,2.

4 2 0

4, 2, 2, 1, 2.

4,4.

4 4 12

4, 4, 4.
5.
5,2.

5, 2, 1.
6.
6,4.

7.
10.

Type.

:;

1

11

2

1
1

5
2

1

109

1

3
1
1

31
1

11
4
4
1
o

18

1

o

3

2

1
3

Var. 1.

Var. 2.

13

Var. 3.

Var. 4.

1
1

10

1

1

is

15

Total Cases.

11

1

24

8
3
1
12
2
1
1
5
2

3
1

146

2

3
1
1

32
1

11
4
4
1
3
1
1
1
1

82
1
3
1
1
1
1
1
8
o

1

5
1

2
1

50 ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA.

As might be expected from the nature of the case, the range of variation of the
syzygial interval for the whole length of the arms is considerably greater than that of
the second interval alone. In the type and in variety 3 it differs also in a tendency to a
decrease rather than an increase in the length of the interval, which is more often two
segments than four, as is seen from Table VI. ; while in the other three varietal spe-
cimens the tendency of the variation is to increase in the length of the interval, four
segments occurring much more commonly than two. "With respect to these varieties,
however, it must be remembered that these conclusions arc all based upon an examina-
tion of single specimens, which, as already mentioned, may in some instances be very
misleading.

(§ 32) The colour of Act. polymorpha is usually (in spirit-specimens) a yellowish
brown, which is much darker in the soft parts of the body than in the elements of the
skeleton. In variety 4, from Ubay, the colour is the same as that of the type, but con-
siderably darker, so that the disk appears almost black. In varieties 1 aud 2 the colour
is rather a greyish brown, which is considerably lighter on the ventral surface of the
disk and arms than on the dorsal skeleton ; and in variety 3 it is a somewhat reddish
brown.

In varieties 2, 3, and 4 the dorsal surface of the skeleton is marked by a median
white line, with more or less defined dark borders, which commences on the radials, and
extends for some distance onto the arms. Its distinctness varies in different specimens
and in different arms of the same specimen ; but it is especially well marked in the
darkly coloured var. 4, as is seen in PL II. fig. 7.

In an adult specimen of Act. polymorpha the total diametex*, including the arms, is
about 200 or 220 millimetres, of which about 20 millims. represents the diameter of the
disk alone ; but in one young specimen I found these two diameters to measure only
105 and 7 millims. respectively. The three specimens of varieties 2, 3, and 4 were of
about the same size as the type, but the single specimen of variety 1 was considerably
smaller, its longer diameter being only about 100 millims., and its shorter (that of the
disk) about 8 millims. This specimen, however, was, I believe, full-grown ; for it had
very large and well-developed genital glands ; while in the young and small specimen of
the type mentioned above, the size of which was about the same as that of variety 1,
the genital glands were scarcely developed at all.

(§ 33) The manuscript name of Act. armata has been given to the type of Actino-
metra here described, by Professor Semper, on account of the small spines with which
the segments of the arms and pinnules are fringed, more especially upon their dorsal
and aboral margins. As, however, this character is a very general one among the Co-
matulce, and as it is by no means so well developed in this type as in many others, I have
thought it advisable not to adopt Professor Semper's specific name, "armata," more
especially as it has been already employed by Pourtales to designate a new American
Anictlon. Under these circumstances, I propose to designate this type as Act. poly-
morpha, having regard to the enormous amount of variation which I have found to exist
in nearly all its characters.

I believe it to be very closely allied to, if not actually identical with, the type described

ME. P. H. CAEPEXTEE ON THE GENUS ACTLNOMETEA. 51

as Alecto parvicirra by Midler1, who gave this name to three small spirit-specimens in
the Paris Museum, from the voyage of Peron and Lesueur in 1803, which I recently
found there under the name of Comatula simplex, Mus. Midler's diagnosis of Alecto
parvicirra was based upon his examination of the three Paris specimens, which all
have an excentric mouth and a terminal comb on the oral pinnules, and is exactly appli-
cable to Act. polymorpha, except that he describes the pinnules as " ziemlich gleichformig."
In their yellow colour and smaller size (about 100 millims.) these also differ slightly
from the type of Act. polymorphic, but without a very much closer examination of them
than I was able to make, it would be impossible to arrive at a definite conclusion as to
the identity or difference of these two species.

The Vavao variety of Alecto parvicirra described by Miiller occurs in the Paris
Museum under the name of Comatula brevicirra, Troschel. This specimen differs from
Act. polymorpha in many subordinate characters, and is not absolutely identical either
with the type or with either of the four varietal specimens which I have examined ;
and I cannot but regard it as representing another of the slight and probably very
numerous modifications of this type, of which I think it most likely that Midler's
original species, Alecto parvicirra, is also a varietal form.

(§ 34) The following diagnosis will, I believe, be found sufficient for the future iden-
tification of Act. polymorpha and of tbe four varieties here referred to.

Actinometra polymorpha, n. sp.

Centrodorsal piece. A circular or irregularly pentagonal disk almost completely concealing the first
radials. Surface flattened, and slightly concave in the centre.

Cirrhi marginal, 15-25, of 11-1-1 segments, of which the fifth and sixth are the longest; hasal ones
thick, and wider than long; remainder taper gradually, and terminal ones are laterally compressed; the
last 5 or 6 segments have a small dorsal spine, increasing in distinctness up to the penultimate segment,
which hears the terminal claw.

Radials 3, of which the first are barely visible ; the second are short, and in the middle of the same
height as the first, but somewhat lower at the sides, for nearly the whole length of which they are united
to one another in pairs.

Axillary radial pentagonal, about twice as wide as the second, to which it is united by ligaments only.

Arms from 13-40 ; rays may divide three times. First segments borne by each axillary in contact
for nearly their whole side.

Distichals and Palmars. When present, 3 ; second bears a long pinnule, and is united to the first
by ligamentous articulation only. Axillary has a syzygium.

Syzygia. First on third brachial, then an interval of 6 segments to the next, and then a general
interval of 3 throughout the arm, variable from 0-6, but usually varying to < 3.

Arms. Anterior arms much longer than the posterior, which are usually non-tentaculiferous. Width
increases from 3rd to 10th segment, remains uniform till about the 14th, and then decreases, slowly
in the long anterior arms, and rapidly in the short posterior ones. Arm-segments wedge-shaped, slightly
overlapping one another, and fringed at the borders with short spines.

Pinnules. The second distichal and the second palmar, when present, bear long pinnules, of which
the palmar bears the shorter one; the next is on the second brachial, and still shorter, and the length

1 ' Gattung Comatula,'' p. 24.
SECOND SERIES. — ZOOLOGY, VOL. II. 8

52 ME. P. H. CABPENTEB ON THE GENUS ACTINOMETBA.

gradually diminishes to the pinnules of the fourth and fifth brachials, -which are the shortest on the
whole arm. From the sixth brachial onwards the pinnules are long and stout, gradually increasing in
length and thickness to near the middle of the arm ; the thickness is greatest in the short posterior arms,
in which both length and thickness rapidly decrease from the middle to the end of the arm, while in the
long anterior arms the thickness slowly diminishes and the length slightly increases, so that the terminal
pinnules are long and slender.

Comb. The last six or eight segments of the distichal, palmar, and first eight or ten brachial pinnules
have the outer ventral margin of each calcareous segment produced into a small lancet-shaped process
which bends over towards the ventral side, so that the end of the pinnule has a comb-like appearance.
Many of the other pinnules till near the end of the arm have similar processes upon their four or five
terminal segments.

Disk. Mouth excentric and interradial ; posterior ambulacra very indistinct, and often nearly obli-
terated. Small calcareous concretions occasionally present in the neighbourhood of the peristome and
anal tube.

Colour. Yellowish brown to dark brown.

Diameter. About 20 centimetres.

Locality. Bohol.

The following are the points in which the varieties differ from the type as described above : —

Variety 1.

Cirrhi. 25, of 13-15 segments, with terminal claw ; spines on dorsal face of terminal segments not very
distinct, but the segments are laterally compressed.

Radials. Second radials completely united with one another in pairs.

Arms. 20.

Syzygial interval. Usually 3, but varying from 1-10 segments ; generally to >3.

Comb. From 2nd distichal to 6th brachial pinnules, and then, at intervals, to about 20th brachial
but no further.

The basal pinnules of the arms have a faint dorsal keel, and the distal ends of their segments are
rather wider than the proximal ends.

Diameter. 105 millims.

Colour. Greyish brown.

Locality. Ubay.

Variety 2.

Centrodorsal piece. Small, but rather thick.

Cirrhi. 10, of 11 or 12 segments, with a terminal claw; the fourth and fifth are longest; the spines
on the dorsal border begin from the middle segments, and the opposing process on the penultimate
segment is well marked.

Radials. Second radials only incompletely united ; second and third very convex, and much higher
than the first ; median dorsal line of skeleton marked by a white line with dark borders, which is lost
about the middle of the arms.

Arms. 29.

Syzygial interval. Usually 3, but varying from 1-10 segments; generally to >3.

Comb. Limited to distichal, palmar, and first five brachial pinnules ; those of the Gth and next
succeeding brachials have a dorsal keel, and the distal ends of their segments are much wider than the
proximal ones.

Colour. Greyish brown.

Locality. Cabulan.

ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA. 53

Variety 3.

Centrodorsal piece large and thick, with only 3 cirrhus-scars.

Radidls. Second radials completely united all round. Centre of dorsal surface of the skeleton, from
the centrodorsal till near the end of the arms, marked by a faint white line with dark borders.

Arms. 39.

Syzygial interval. Usually 3, but varying from 1-7 segments; generally to <3.

Comb. On second distiehal, palmar, and brachial pinnules, and occasionally also on those of the
3rd— 5th brachials, but on no others.

Colour. Reddish brown.

Locality. Bohol.

Variety 4.

Centrodorsal piece large and thick, with only 3 cirrhus-scars.

Radials. Second radials closely united all round. Median white line on dorsal surface of skeleton
very marked.

Arms. 33, all tentaculiferous, and tolerably uniform in length and in the character of their pinnules.

Syzygial interval. Usually 3, but varying from 0-6 segments; generally to >3.

Pinnules. Oral pinnules much stouter than in the type ; that of third brachial but little shorter than
that of second. Comb limited to these and to the distiehal and palmar pinnules, and the processes forming
it gradually come to rise from the ventral surfaces of the calcareous segments instead of from their outer
margins.

Colour. Blackish brown.

Locality. Ubay.

IV. Tiie Skeleton.
(i.) The Skeleton generally with its Ligaments and Muscles.

(§ 35) The general structure of the skeleton of Actinometra, and of the ligaments
and muscles which connect its component pieces, is precisely the same as in Antedon ;
and as this has been already described by Dr. Carpenter 1, there is no need to repeat it
here : a few points, however, must be treated somewhat more in detail. The component
pieces of the skeleton of Actinometra, as of all the other Echinoderms, consist of a
calcareous reticulation formed by the calcification of an organic basis of a protoplasmic
nature, in which numerous nuclei and pigment-granules are imbedded. This " nuclear
tissue," as Simroth2 has called it, is in the form of a network, around the meshes of
which the calcareous material is deposited.

The character of the calcareous reticulation varies greatly in different parts of the
skeleton, being much closer at the synostoses and syzygia and at the articular surfaces
than in the interior of the segments ; and in correspondence with this greater compact-
ness of the calcareous tissue, the organic plexus which forms its basis becomes remark-
ably modified at these points, as will be seen further on. The various modes of union
of the different pieces of the Crinoid skeleton have been closely investigated by Miiller
and by Dr. Carpenter. The former3 described the stein-segments of Pentacrinus as united
to one another in two different ways — (1) by the tendons which traverse the whole

1 Phil. Trans, loc. cit. p. 702. : Op. cii.p. 433. 3 Bau des Pentaerinus, pp. 17-20.

54 ME, P. H. CARPENTER OX THE GENUS ACTIXOMETRA.

length of the stem, passing through the substance of its various segments, and (2) by the
" elastic interarticular substance" between the individual segments.

The substance of these tendons consists of a white fibrillar tissue very like the ten-
dinous tissue of the higher animals ; but Midler supposed the elastic interarticular sub-
stance to be of a totally different nature, consisting of " lauter senkreckt stehenden
Fasersaulchen, die durch Reihen bogenformiger Schlingen einfacher Fasern verbunden
sind," and " diese Scblingen gehen mit den regelmiissigsten Arkaden in ganz gleichen
Abstiinden aus einem Fasersaulchen in das andere liber." This substance fills up
the whole space between the successive stem-segments which is not occupied by the
tendons, and is connected in the closest possible manner with the opposed surfaces of
every pair of segments, even extending for a short distance into their superficial cal-
careous tissue. Each of the arcades above mentioned consists of a single primitive
fibre, the terminations of which are lost in the " Fasersaulchen ;" and the passage of
these fibres in loops from one fibrous column to another gives an elasticity to the whole
tissue, viz. a power of contraction after lateral displacement, and of extension after
vertical compression, although the individual fibres are not elastic in the ordinary
sense of the word.

Miiller described the basals of Pentacrinus as united with the top stem-segments in
the same manner as the successive stem-segments Avith one another, namely, by this
elastic interarticular substance, while their sides simply " stossen an einander " (p. 25).
This mode of union between the stem-segments was generally called by him a "Nath,"
or suture ; but he sometimes spoke of it as an articulation, though he usually employed
this last term only in those cases in which two segments are movable on one another
through the intervention of muscles and ligaments which pass between them.

He further described the union between the first radials and the basals of Pentacrinus
and between the first radials and the centrodorsal piece of Comatula as a suture, which
name he also gave to the lateral union of the five first radials with one another (pp. 28, 29) ;
but he does not seem to have supposed that in these cases the various elements were
connected by the elastic interarticular substance which he found between the likewise
suturally united stem-segments. In fact, in speaking of the syzygia, which he called an
immovable sutural union of two segments, he said expressly that not only the muscles
but also the clastic interarticular substance was absent, On the other hand, the latter is
to be found between the segments which are capable of motion upon one another, whether
ligaments and muscles be present, as between the first and second radials and between
most of the brachials, or ligaments onlv, as between the first and second brachials and the
second and third radials ; for Miiller described the ligaments connecting two mutually
movable segments as having essentially the same structure as the elastic interarticular
substance of the stem, except that their surface is plain and not " krausenartig gefaltet"
(pp. 30-38).

In those more common cases in which there is a muscular union between two segments,
such as the first and second radials, in contact by transverse articular ridges upon their
opposed faces (PL VII. figs. 1 b, 2 a, I b, 5 a ; I), Miiller drew no distinction between the
pair of ligamentous bundles on the ventral side of the articular ridge, and the single

MR. P. H. CARPENTER ON THE GENUS ACTINOMETRA. 55

mass which occupies the whole space between the opposed faces on the dorsal side of
this ridge, describing them as alike consisting of elastic interarticular substance, the
function of which is extensor and antagonistic to the flexor action of the muscles.
Dr. Carpenter !, however, regards the former as interarticular, with the special function
of holding the pieces together, but allowing a certain amount of movement between
them, while he describes the single dorsal mass as elastic, and as antagonizing by its
extensile powers the action of the flexor muscles. Histologically he finds no differ-
ence between them, both consisting of minute, straight, and nearly parallel fibres, very
much, in fact, like those which Miiller described as composing the tendons of the stem of
Pentacrinus. At their points of attachment to the pieces of the skeleton these fibres
pass into their basis substance and become incorjiorated with it.

The union of the first radials with one another and with the centrodorsal piece, which
had been spoken of as sutural by Miiller, is described by Dr. Carpenter as " an
adhesion of expanded surfaces closely fitted together, and held together by the con-
tinuity of their sarcodic basis substance " (p. 701) ; so that the different elements are
cemented together by a " thin layer of sarcodic substance, continuous with that which
occupies the meshwork of their own calcareous reticulation." This mode of union may
be conveniently described by the word " synostosis," which has been employed by
Simroth to designate the mode of union of two faces which " verkitten sich " in the
skeleton of the Ophiuridce. It is essentially the same as the syzygial union which
occurs between certain pairs of the primitive arm-segments, although differing from it in
points of detail.

(§ 3G) We have seen that Miiller regarded the tendinous tissue of the stem of 'Penta-
crinus, and the fibrous ligamentous bundles, or, as he called it, the elastic interarticular
substance uniting the movable elements of the skeleton both of Pentacrinus and of
Comatula, as distinct from one another. I believe, however, that they are fundamentally
identical, not only with one another, but also with the so-called "cement-substance"
between two segments which are united by synostosis. This last consists, in the Ophin-
ridce, according to Simroth 2, of connective-tissue fibres which lose themselves in the
organic basis of the skeleton, and are of the same nature as the substance of the
masses of connective tissue uniting two articulating surfaces, both tissues staining
deeply with picro-carmine. I find the same to be the case in Comatula and Pentacrinus.
The tendons of the stem of the latter genus, the ligamentous bundles, composed, according
to Miiller, of elastic interarticular substance, which connect every pair of movable
arm-segments, and, lastly, the "cement-substance," uniting the first radials to one
another and to the centrodorsal piece, all stain very deeply with picro-carmine, and are
of essentially the same histological structure.

Dig. 1, on Plate III., represents a portion of a horizontal section through the suture,
or, as I prefer to call it, the synostosis of two of the first radials of Pentacrinus. In the
immediate neighbourhood of their apposed lateral faces there are none of the nuclei nor
pigment-granules which are imbedded so abundantly in the more internal portions of their
protoplasmic ground-substance, and the threads of the plexus of which it is composed
1 Phil. Trans, he. cit. pp. 703-714. - Op. cit. p. 435.

56 JIE, P. H. CAEPEXTER OX THE GENUS ACTIXO.METEA.

become excessively attenuated and disposed with great regularity almost parallel to one
another. At the same time the meshes of this organic plexus become greatly elongated
in the intervals between the parallel threads or fihres, which are connected with one
another by very delicate fibrils passing in the form of loops from one fibre to another.
These loops, which forcibly recall Midler's description of the arcades connecting the
fibrous columns of the elastic interarticular substance in the stem of Pentacrinus, are
simply the expression of the ends of elongated meshes of the protoplasmic plexus
forming the organic basis of the skeleton. In the neighbourhood of each of the two
opposed surfaces the fibrous elements of this plexus assume the character of closely
placed parallel connective-tissue fibres, with no pigment-granules nor nuclei imbed-
ded in them, but staining deeply with picro-carmine, while the normal protoplasmic
basis of the interior part of the calcareous segments is but little affected by this
reagent. These fibres pass from the organic basis of the one segment into that of the
other so that the two are firmly united, and the superficial denser layer of calcareous
tissue is deposited around their ends, which corresponds with Midler's description of the
elastic interarticular substance of the stem of Pentacrinus as extending for a short
distance into the calcareous substance of the opposed faces of the segments. The
superficial layer of calcareous reticulation which occupies the small intervals between
the ends of the fibres thus becomes extremely close and compact ; but the central
portion of the fibrous tissue (PI. III. fig. 4, L) does not calcify, remaining as a thin layer
of fibrous cement-substance between the two opposed surfaces, precisely like the inter-
articular substance in the stem of Pentacrinus, with which I believe it to be identical.
It is, at any rate, of the same nature as the substance of the ligaments connecting the
first and second radials, which Miiller described as identical with that connecting the
stem-segments ; for at the angles of the radial pentagon the fibres of the cement-sub-
stance connecting the adjacent first radials with one another in pairs pass directly into
the fibres of the ligamentous bundles between the first and second radials. These, which
are of precisely the same character as the ligamentous bundles between the successive
brachial segments, also stain deeply with picro-carmine, and only differ from the cement-
substance in the greater length of their fibrous element.

At the points of attachment of the ligaments to the pieces of the skeleton, the meshes
of the organic plexus become greatly elongated, and its fibrous bars regularly disposed
and connected with one another by loops, as above described. As, however, the distance
between the two articulating faces is very much greater than in a synostosis, several of
these minute primitive fibrils unite to form one of the larger fibres composing the liga-
mentous bundle, at the other attachment of which these primitive fibres again separate,
become connected with one another by transverse loops, and finally pass into the bars
of the protoplasmic plexus forming the ground-substance of the next segment.

(§ 37) I have found the fibres composing the ligamentous bundles between the
arm-segments of Antedon to terminate in the manner above described for Pentacrinus;
but in Act. poli/morpha they do not pass so directly into the organic basis of the seg-
ment. At the ends of the ligamentous bundles, where their component fibres begin to
break up into primitive fibrils, the latter cross one another in all directions, very much

ME, P. H. CAEPENTEE OX THE GENUS ACTINOMETEA. 57

as described by Simrotb 1 in Ophiactis virens, so as to form a network of delicate
threads without any imbedded nuclei, although it may contain pigment-granules ; and
this network passes very gradually into the nucleated protoplasmic plexus forming the
organic basis of the brachial segments (PI. III. fig. 7, Lx).

The tendons of the stem of Pentacrinus are, I believe, of precisely the same character
as the ligamentous bundles between the arm-segments, although, of course, enormously
longer. They stain deeply with picro-carmine, and are composed of parallel fibres, which
may be teased out into very much finer ones, and their upper ends pass into the organic
ground-substance of the five basals, precisely in the same manner as the fibres of the
arm-ligaments pass into the protoplasmic network composing the organic basis of the
successive segments.

In Pentacrinus Wyville-Thomsoni, in which the five basals are completely in contact
with one another in pairs, the two elements of every pair are united by a synostosis,
and the union of the basals with the radial pentagon above them is of the same cha-
racter. The first radials of Comatula are connected with one another and with the
centrodorsal piece in the same manner, as is seen in PI. III. figs. 5, 6, where L, I repre-
sent the tracts of fibrous tissue connecting the first radials with one another and with
the centrodorsal piece respectively. The terminal portions of this fibrous tissue become
calcified to form the compact superficial layers of calcareous substance on the apposed
faces, while the middle portion remains as the fibrous cement-substance uniting the two
calcareous segments, which is thus essentially of a connective-tissue nature.

The mode of union of the segments of the calyx of the Tesselate Crinoids, none of
which are connected with one another by a muscular articulation like the first and
second radials of Pentacrinus and Comatula, was most probably a synostosis of the
same nature as those just described. The immovable sutural unions between certain of
the brachial segments to which Miiller gave the name of " syzygia," arc, in Pentacrinus,
of precisely the same nature as the synostoses between the segments of the calyx, the
organic basis of the one segment being continuous with that of the other through the
fibrous cement-substance, which forms a thin layer between the whole of the two simple
opposed surfaces. This was described by Miiller2 as a very delicate membrane, of a
different nature from the elastic interarticular substance between the likewise sutu-
rally united stem-segments.

In Comatula, however, the apposed surfaces of the two elements united by a syzygium
are not plain and simple, as in Pentacrinus and Bhizocrinus3, but marked by a series of
radiating ridges, as in Apiocrinus obconicus, Goldf. 4 The ridges of the two surfaces
correspond in position, and when the surfaces are in contact are closely applied to one
another, and united by fibrous cement-substance as in an ordinary synostosis. The fibrils
are very numerous and placed very close to one another, so that the calcareous reti-
culation forming the ridges is remarkably dense and compact, being formed around the
ends of these fibrils where they pass into the organic basis of the segments ; and these
ridges thus correspond to the whole of the syzygial surfaces in Pentacrinus and Rhizo-

1 Op. cit. p. 435. - Eau des Pentacrinus, p. 20. 3 Sars, he. cit. p. 22. " Petref. German. Taf. lvii. fig. 5.

58 ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA.

crmus. In the intervals between them the organic basis of the one segment is directly
continuous with that of the other, little or no fibrous tissue being interposed.

The muscular fibres of Actinometra correspond very closely with those of Antedon, as
described by Dr. Carpenter l and Ludwig 2 ; their expanded terminations are simply
applied to the surfaces of the calcareous segments to which they are attached, not passing
into their substance as the ligamentous fibres do (PL III. fig. 7, b ms), and there is no
trace either of a sarcolemma or of transverse striation.

(ii.) The Dorsal Cirrlii.

(§ 38) The Dorsal Cirrhi of Actinometra do not appear, so far, at least, as my observa-
tions have extended, to be developed over such a large surface of the centrodorsal piece
as is the casein A ntedon. In all the specimens which I have examined the cirrhi are
limited to its margin, while its central portion is entirely free from them and usually
slightly concave. There is generally only one row of these appendages ; but small and
rudimentary ones may occasionally be found interposed between the large and full-grown
ones at the extreme circumference of the plate, thus forming the commencement of a
second row. The number of cirrhi existing at any one time upon the plate-like centro-
dorsal piece of Act. polymorpha varies, I believe, between 15 and 20. Three or, in large
specimens (PL VI. figs. 1, 2), four are attached on each side of its more or less distinctly
pentagonal margin, while in var. 1 (PL VI. fig. 14), and in one specimen of the type
(fig. 7), the total number reached 25. In var. 2 (fig. 16) there are only 10; while
in vars. 3 (fig. 20) and 4 there is no evidence, in the single specimens which I have
examined, of the existence of more than three perfect cirrhi in the adult state, as there
are no sockets around the margins of the centrodorsal plate for the attachment of a
larger number ; though there may be minute openings here and there, which appear to
have corresponded with the central canals of lost cirrhi, the sockets of which have been
obliterated by a later calcareous deposit.

It is not a little singular that the dorsal cirrhi of Act. polymorpha, like the centro-
dorsal piece which bears them, should exhibit such a very slight range of variation, not
only in size but also in number (the three varieties just mentioned of course excepted) ;
for in nearly every other part of the skeleton the range of variation is very great. In
Antedon rosacea the reverse appears to be the case ; for the composition of the skeleton
is fairly constant in its simplicity, but the cirrhi vary considerably both in number
and in size.

In a fully developed cirrhus of Act. polymorpha (PL III. fig. 8 a) the number of
segments varies from 11-11, being usually 12 or 13, the last of which is in the form of
a stroug sharp claw. This is attached by simple suture to the penultimate segment,
which is prolonged at the base of the claw into a short opposing process on its concave
or aboral margin.

The diameter of the basal segment somewhat exceeds its length ; but in the second and
third segments this disproportion between the length and breadth is reduced, and in the
fourth it becomes reversed, the length of this segment being slightly greater than its

1 rhil. Trans, he. cit. p. 704. 2 Beitrage <fcc. he. cit. p. 40.

ME. P. H. CAEPEXTEE OX THE GEXTJS ACTIXOMETBA. 59

diameter. In the fifth and sixth the proportion between the length and breadth reaches
3:2; and a very slight degree of lateral compression is visible in the latter segment,
Avhile the fifth, like the four basal segments, is cylindrical, or nearly so. These are the
two longest segments of the cirrhus ; and from this point onwards the length of the
segments gradually decreases, until in the tenth and following segments ft becomes
again less than the dorso-ventral diameter. At the same time the transverse diameter,
which in the first five cylindrical segments is equal to the dorso-ventral diameter, under-
goes in the seventh and eighth segments a sudden decrease. A faint indication of this
is seen in the sixth segment, and it is continued on to the end of the cirrhus, so that its
terminal portion exhibits a considerable degree of lateral compression. In correspond-
ence with this, a small spine gradually developes itself on the dorsal margins of the sixth
and successive segments, on which it becomes progressively more and more marked,
until in the penultimate segment, i, e. the last one before the claw, it becomes the short
pointed opposing process above mentioned. This series of small spines, like the single
penultimate opposing process of the cirrhi of Ant. rosacea, seems to be characteristic of
those cirrhi only which have reached their full development ; for scarcely any trace of it
is visible in the still immature cirrhus represented in PI. III. fig. 8 c. Although its
penultimate segment shows a faint indication of an opposing process, the dorsal margins
of the segments immediately preceding it are almost, if not quite, even.

The segments of this cirrhus, although of the normal proportions, are not so large as
those of an adult cirrhus, and (counting the terminal claw) are II in number, the first
seven of which precisely resemble those of the cirrhus represented in PI. III. fig. 8 a,
with 12 segments and a chaw, so that the extra one in the former case would appear to
be interpolated between the middle and end of the cirrhus. As in Ant. rosacea, and for
precisely the same reason, greater facility of flexure, the ligamentous substance between
the terminal segments is thicker on the aboral side than on the oral (PI. III. ficr. 9 b),
while in the more cylindrical segments of the basal half of the cirrhus the ligamentous
substance is tolerably equally developed on both sides (fig. 9 a). In correspondence with
this, the canal {cc) which occupies the centre, or nearly so, of the circular opposed
faces of the basal segments (fig. 9 a), lies in the laterally compressed terminal segments
much nearer to the oral side of the oval articular faces, more than half of which is
occupied by the large fossa for the lodgment of the aboral interarticular ligament
(fig. 9 b). The opening of the central canal (c c) is surrounded in each case by a more or
less prominent articular surface.

(§ 39) Both the single specimens of Varieties 3 and I of Act. pohjnwrpha which
I have been able to examine had unfortunately lost all their few cirrhi ; but in Var. 1
twenty-five still remained attached to the centro-dorsal piece ; most of these are fully
developed, and present some slight differences from those of the type (PI. III. fig. 11).
Not only is the number of segments greater, varying usually from 13 to 15, besides the
terminal claw, but they also differ considerably in their relative proportions ; for in the type
the fifth and sixth segments are the longest, while in Var. 1 there is less difference between
them and the fourth and seventh in this respect. The lateral compression, which is not
visible till the eighth segment, becomes somewhat marked towards the end of the cirrhus,

SECOND SERIES. — ZOOLOGY, VOL. II. 9

60 ME. P. H. CARPENTER ON THE GENUS ACTINOMETEA.

which is more distinctly flattened than in the type, although the small opposing spines
on the dorsal margins do not appear at all until the three or four penultimate segments ;
and even on these they are but slightly developed.

In Var. 2 there are only ten cirrhus-sockets around the margin of the pentagonal
centrodorsal plate, two on each of its sides, and placed close to the angles (PL VI. fig. 16).
The number of segments in each cirrhus is eleven or twelve besides the terminal claw
(PI. III. fig. 10), and the width of the two basal segments somewhat exceeds their
length. In the third the length and breadth are nearly equal, but in the fourth and
fifth, which are the two longest segments of the cirrhus, the proportion between them
is as about 4 to 3. The sixth segment is slightly shorter than the fourth, and from this
point to the end of the cirrhus the length of the segments gradually decreases, while at
the same time they exhibit a slight degree of lateral compression. The dorsal spine,
the first indication of which is seen in the fifth and sixth segments, becomes very
marked indeed towards the end of the cirrhus, and develops in the penultimate segment
into a stout opposing process.

(§ 10) The development of the cirrhi of Act. polymorpha seems to take place somewhat
differently from their development in Aid. rosacea as described by Dr. Carpenter L. In
the latter species the individual segments usually present all the characters of maturity
from a very early date, viz. the relative proportions in the length and breadth of the
segments, the bevelling off of the opposed faces on the aboral side, and the develop-
ment of the opposing process on the penultimate segment. But in some rare cases, even
after the cirrhus has attained a considerable size and has the normal number of seg-
ments, the latter are of a very rudimentary character ; their basal segments are the
longest, and the following ones rapidly decrease in diameter, so that the whole cirrhus
tapers considerably from its base to its point. This condition gradually becomes less
and less marked as the segments increase in size, and their opposed faces become
bevelled off towards the aboral side, so that the cirrhus ultimately acquires all the
characters of maturity.

This mode of development, which is the exception in Ant. rosacea, seems to be the
rule in Act. j}olijmor})ha. All the very young cirrhi, both of the type and of Varieties
1 and 2, which I have met with, taper rapidly from the base to the apex ; and while
the four or five basal segments exhibit from a very early jieriod the same proportion
between length and diameter as is seen in a completely developed cirrhus, the following
ones are still in a very rudimentary condition. The sixth segment, instead of being as
long as the fifth, is much shorter ; the seventh is still shorter and more slender, while
the terminal segments are little more than a succession of small disks ending in a small
and very rudimentary claw (PI. III. fig. 8 b). They are thus not only of the smallest
dimensions, but have a much more immature appearance than the basal segments; and
it would therefore seem that the augmentation in the number of segments is effected
by the interpolation of new segments, not at the base, as is usually the case in Ant
rosacea, but between the middle segments and the terminal claw.

1 Phil. Trans, he. tit. p. 711.

ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA. 61

The fact already mentioned (sect. 38), that in two cirrhi in which the number of
segments is different, the character of the first six or seven is the same, would seem to
point to the same conclusion.

(iii.) The Centrodorsal Plate.

(§ 41) In all the Actluometrm with which I am acquainted, the external appearance
of the " Knopf," or centrodorsal piece, is very characteristic. Like the cirrhi which it
bears, it is far more constant throughout a considerable range of species from very
various localities, than it appears to be in the individual members of a single species,
both of the European and of some of the foreign Antedons, even when existing in the
same locality.

Thus, for example, the centrodorsal piece of Ant. celtiea may either be very shallow,
flattened, and bluntly rounded off at the base, where it was originally united to the
joint of the stem next beneath it, with only two rows of sockets ( V) for the attachment
of the cirrhi (PI. IV. fig. 1) ; or it may be deep and nearly hemispherical, with three or
four alternating rows of cirrhus-sockets, and terminating inferiorly in a flattened circular
or rudely pentagonal base (fig. 6) ; or, finally, it may have the form of a five-sided
pyramid, the apex of which is directed downwards and very slightly truncated, while
the sides bear three or four alternating rows of indistinct cirrhus-sockets (PI. IV. fig. 8).
This last character indicates that the animal had attained a considerable age ; and as all
the specimens of this kind which I have seen have been smaller than normal specimens
of Ant. celtiea, and exhibit slight differences from them in the characters of other parts
of the skeleton, I am disposed to regard them as dwarfed varietal forms rather than as
young and incompletely developed individuals of the ordinary type.

Again, of the two specimens of Ant. macrocnema, Val., in the Paris Museum, both of
which were brought from New Holland by Quoy and Gaimard in 1829, one has a large
and hemispherical centrodorsal piece, while in the other it is a short pentagonal or
nearly circular column, on which the cirrhi are disposed in three or four alternating
rows.

These instances, which might be greatly multiplied, suffice to show that the centro-
dorsal piece of Antedon may vary very considerably in its external appearance. In Acti-
nometra, so far as my experience goes, it is almost invariably a flattened circular, or
rudely pentagonal disk, somewhat hollowed in the centre of its dorsal surface, and with
low sloping sides marked out into distinct sockets for the articulation of the cirrhi
(PL V. figs. 1, 6, 15, and PL VI. figs. 1, 2, 7, 14, 16, 20). It generally conceals more or
less of the first radials which rest upon it (PL II. figs. 9-11, cd), but it may sometimes
be very irregularly extended so as to conceal a considerable part of one or more of the
second radials (PL II. fig. 8, c d).

As a general rule, only one row of cirrhus-sockets can be traced ; but in the large
Act. robusta (PL V. fig. 15) I have found two alternating rows of sockets to be distinctly
visible, and even traces of a third row, so that the dorsal surface of the plate becomes
dightly convex, though by no means to the same extent as in most Antedons. In fact,
the flattened plate-like condition of the centrodorsal piece, and the existence upon

9*

62 ME. P. H. CAEPEXTEE ON THE GENUS ACTINOMETEA.

it of only one or two whorls of marginal cirrhi, seem to be very characteristic of Actino-
metra ; and it would be almost possible to distinguish the Antedons from the Actino-
metrce among the Comatulce described in Midler's memoir, by simply referring to his
descriptions of the " Knopf."

When the centrodorsal is viewed in situ, with all the cirrhi attached around its sides,
its central flattened surface appears almost circular ; but when the cirrhi are removed,
so as to expose the low and more or less sloping sides to which they are attached, the
outer margin of the plate is seen to have a distinctly pentagonal form. This is well
seen in the large Act. robusta (PI. V. fig. 15), with its three rows of cirrhus-sockets ; but
in Act. Solaris (PL V. fig. 1) the angles of the pentagon are more rounded off, and there
is only one row, a complete one, however, of cirrhus-sockets, while the plate itself is of
a considerable diameter, so as entirely to conceal the first radials.

In a variety of this species, which I believe to be identical with the Act. pectinata of
Hetzius and Midler, the diameter of the centrodorsal (PL V. fig. G) is very slight, so that
the greater part of the superjacent radial pentagon is visible outside its pentagonal
margin ; and there are only ten distinct cirrhus-sockets, two at each angle, though one
of the angles (the upper one in the figure) is marked by the presence of a third socket,
which either indicates the commencement of a second row of cirrhi, or, and more pro-
bably, tbe incomplete obliteration of a pre-existing row.

In Act. polynwrpha the angles of the pentagonal centrodorsal plate, the diameter of
which varies from 3 to 5 millims., are sharp and distinct (PL VI. fig. 2) ; in varieties 3
(fig. 20) and 4, the dorsal and ventral surfaces almost meet at tbe edge of the plate,
which is very slightly truncated, and marked in three places by the large sockets ( U)
for the attachment of the few remaining cirrhi, while other small openings indicate the
former existence of others which have been since lost.

In var. 2 (PL VI. fig. 16) the diameter of the centrodorsal plate is very small, as in
Act. pectinata (PL V. fig. 6) ; and, also as in this species, it normally bears ten cirrhi, two
at each angle. The sockets for the cirrhi are, however, but slightly marked, so that the
edge of the plate is but little truncated, and it can hardly be said to have distinct sides
as in Act. Solaris (PL V. fig. 1).

In the type, on the other hand (PL VI. figs. 1, 2, 7), and more especially in var. 1
(fig. 14) j in which there are 25 cirrhi, the edge of the plate where the dorsal surface
passes over into the ventral is obliquely truncated, so that the plate has distinct sides,
which are marked by numerous cirrhus-sockets, as in Act. Solaris (PL V. fig. 1). Towards
the ventral surface the angles of the pentagonal margin of the plate are prolonged into
five short processes (PL VI. fig. 14, t), each situated between a pair of cirrhus-sockets (U).
Their distinctness varies in different individuals ; they are especially marked in var. 1
(fig. 14), and in that specimen of the type which resembles it in having 25 cirrhi (PL VI.
fig. 7), and they are best seen after removal of the centrodorsal plate from the radial
pentagon which rests upon it. They exist also in the other varieties of Act. polymorpha
(PL VI. figs. 16, 20), though they are not so distinctly visible externally, owing to the
greater extension of the dorsal surface of the plate than is found in the type and in

ME. P. H. CAEPENTEE ON THE GENUS ACTLVOMETEA. 63

var. 1, since it is not reduced in size by the truncation of its edges, as is the case in
these forms.

(§ 42) The meaning of these processes becomes apparent when we examine the
ventral or superior surface of the centrodorsal plate which bears the superjacent radial
pentagon. Its condition in Act. polymorpha is rather complicated, and will be better
understood after an examination of the simpler forms of this surface presented by other
species of Comatula.

In the large variety of Ant. celtica, the cavity of the centrodorsal piece (PL IV.
fig. 2, cd.c) is very deep, and its opening on the flattened ventral surface has a circular
or somewhat pentagonal form. Between the angles of this internal pentagon and those
of the outer more distinctly pentagonal margin of the piece run five slight ridges or
elevations (i.e). In the intervening radial areas (r.ar), between these ridges the surface
is somewhat depressed, so as to receive the convexities of the dorsal surfaces of the first
radials that rest upon it, while the five ridges correspond with five slight furrows
marking the lines of junction of these surfaces on the dorsal aspect of the radial pen-
tagon (PL IV. fig. 3), and are therefore interradial in position.

In Ant. celtica these ridges are tolerably uniform in width throughout their whole
course ; but in Ant. rosacea (PL IV. fig. 15) they are considerably wider at their internal
or central ends than they are towards the external pentagonal margin of the piece, so
that they have an elongated triangular form. Prom the base of each triangle a shallow
depression extends a little way towards its apex, cut out along the median line of the
ridge ; but it soon becomes obliterated by the gradual approximation of its two sides,
which meet and form a simple ridge, like that of Ant. celtica (PL IV. fig. 2), extending
outwards to the margin of the plate.

The central ends of the radial areas into which the ventral surface of the plate is
divided are marked in Ant. rosacea by five shallow depressions (PL IV. fig. 15, q),
placed close to the margin of the internal cavity, which bends somewhat inwards at
these points. These depressions correspond in position with five large radial openings
on the dorsal surface of the pentagonal base of the calyx (PL IV. fig. 16, Q), and receive
the blind euds of five diverticula of the body-cavity, which are enclosed between the five
radial spout-like processes of the rosette (PL IV. figs. 13, 16, />) and the internal faces
of the first radials. They are, however, occasionally absent in Ant. rosacea, whde, on the
other hand, traces of them may occur in Ant. celtica. In fact, the differences which
I have described above in the appearance of the ventral surface of the centrodorsal piece
of these two species must not be regarded as representing more than two extreme
variations of one and the same type.

(§ 13) We shall now be able to understand the meaning of the short processes (t)
above mentioned, which are seen projecting from the angles of the centrodorsal plate of
Act. polymorpha, when viewed from below. It will behest to begin with the examina-
tion of the ventral surface of the centrodorsal piece of variety 1 (PL VI. fig. 15), in
which they are more distinctly marked than in the type. This surface rises slightly
from the circumference towards the centre, which is occupied by the opening of a
shallow cavity, the centrodorsal ccelom (cd.c), the diameter of which is rather less than

64 ME. P. H. CAEPENTER OX THE GENUS ACTINOMETRA.

one third of the total diameter of the plate. The floor of this cavity is marked hy minute
punctations (u), which are the internal orifices of canals proceeding from it towards
the dorsal surface of the plate. They originally opened externally on the summits of
the small tubercles occupying the centres of the sockets for the articulation of the
first developed cirrhi in the young animal ; hut their openings have gradually become
obliterated by the deposit of new material upon the central portion of the external
surface of the plate, as described by Dr. Carpenter ' in Antedon rosacea. This is
accompanied by the continual removal of old material from the internal surface, so that
the minute openings (PL IV. fig. 15, u) seen on the central part of the floor of the
internal cavity of the centrodorsal piece are the original external openings of the first
developed canals, which have subsequently become closed externally by the new material
deposited upon the central part of the dorsal surface. The internal openings of the
canals proceeding to the last developed cirrhi are much larger, and placed more towards
the periphery of the floor of the cavity. Similarly in Actlnomelra polymorpha, the
internal openings of the canals proceeding to the existing marginal cirrhi on the plate-
like centrodorsal piece are placed under its projecting rim, so as not to be visible from
above. There are usually one or two large openings under the central margins of each
of the radial areas (PI. VI. figs. 3, 8, 10, 15, 17, 21, r.ar), and the canals which proceed
outwards from these internal openings break up into five branches, one of which reaches
the summit of each of the small tubercles occupying the centres of the five cirrkus-
sockets, which are placed along the outer or peripheral margin of each of the corre-
sponding radial areas (PL VI. figs. 1, 2, 7, 14, 16, 20, 77). These canals enclose the axial
cords of the cirrhi (PL VIII. figs. 1, 3, 4, 5, G, 8, n.c), which proceed from the fibrillar
envelope of the quinquelocular organ contained in the cavity of the centrodorsal plate
(PL VIII. figs. 1, 2, 3, 7, -V), and surround the cirrhus vessels arising from its chambers
(figs. 2, 3, 7, ch), from each of which there arises a single trunk ", dividing, sooner or
later, into branches for the individual cirrhi. In the specimen of var. 1 represented in
PL VI. fig. 15, the division is not completed within the cavity of the centrodorsal plate,
as two or, sometimes, even only one aperture can be seen under the inner margin of
each of the radial areas, so that the primary trunk enters the substance of the plate, and
there divides into the five branches for the cirrhi placed on the outer margin of each
radial area.

The rim of the cavity of the centrodorsal plate of Actinometra polymorpha, var. 1
(PL VI. fig. 15, cd.c), is ten-sided, or nearly circular, and is not marked by shallow
radial depressions, like those described above in Ant. rosacea (PL IV. fig. 15, q). The
radial areas rise very slightly from their peripheral to their central margins, and are
marked by various indistinct ridges and furrows. Their sides rise towards the five inter-
radial elevations, which, though not very much raised above the general surface of the
plate, are nevertheless very distinct ; for they are wide and marked by shallow grooves

1 Phil. Trans, lot. tit. pp. 742, 743.

2 This is in precise accordance with the origin of the vessels proceeding to the cirrhi which arc borne on tho stem
of Pen.tacrinus. At every nodal segment the five chambers which are placed radially around the central axis of the
stem enlarge slightly, and each gives off a single vessel to one of the five cirrhi.

ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA. 65

(PL VI. fig. 15, b.g), which occupy the greater part of their width, so that the simple
ridge, as seen in Ant. celtica (PI. IV. fig. 2, i.e), is here represented by the two sides of
the groove which is cut out along its median line. In Ant. rosacea (PL IV. fig. 15), as
we have already seen, these sides meet at a very short distance from the central end of
the groove so as to obliterate it. In this form, however, they approach one another very
gradually, and only just meet within the margin of the plate ; but the ridge formed by
their fusion does not end here as in Ant. rosacea, for it is continued a short distance
beyond the general surface of the plate, so as to appear as a short process (t) extending
outwards from the angle between two sides of its external pentagonal margin. Conse°-
quently these five short processes appear on the dorsal aspect of the plate, prolonging its
angles outwards, as we have seen in sect. 11 (PL VI. fig. 11, t).

(§ 14) The grooves (b.g) which are thus cut out along the median line of the inter-
radial elevations on the ventral surface of the centrodorsal plate of Actmometra are of
no little importance ; for there lie in them, as will be seen further on, the five rays of the
basal star (PL VI. fig. 13, S), which is in close connexion with the dorsal surface of the
radial pentagon; they may therefore be called the " basal grooves " (b.g).

As a general rule, these interradial elevations and the basal grooves are, like the
rays of the basal star, entirely devoid of pigment, which is, however, very abundant in
the organic basis of the calcareous reticulation composing the rest of the ventral surface
of the plate ; so that when this is first exposed by the removal of the centrodorsal from
the dorsal surface of the radial pentagon which rests upon it, five white rays are visible
on a dark background. Unless the plate is immediately removed from the alkaline
solution used to effect its separation, this distinction iu colour between the radial and
the interradial portions of its ventral surface rapidly disappears, owing to the destruc-
tion of the pigments contained in the former.

The development of these basal grooves is not only different in the type and in all the
varieties of Act. polymorphs but it differs in different individuals of the type, and even
to a certain extent in the same individual.

In the specimen of the type represented in PL VI. figs. 7, 8, which, like variety 1,
had 25 cirrhi, two only of the basal grooves are seen ; for the other three are occupied by
the rays of the basal star (fig. 8, S), which have become detached from the rest of the
star and from the first radials with which they were connected. But even these two
grooves do not resemble one another ; one extends almost to the margin of the plate,
beyond which the interradial ridge formed by the union of its sides is continued as a
short process (t), just as in var. 1 (fig. 15). The other open groove, however, terminates
very soon, as its sides, widely separated at its central end, bend sharply towards one
another, and meet some distance within the margin of the plate, to which the ridge
formed by their union does not extend, for it terminates abruptly in a blind and
rounded extremity.

In variety 1 (PL VI. fig. 15) the basal grooves (b.g) are narrow, and after increasing
a little at first, diminish gradually in width from their central to their peripheral ends°;
but in the specimen of the type (fig. 8) they are much wider in proportion to their
length, and the width increases slightly from their base to about the middle of their

66 iCE. P. H. CAEPZVTEE OX THE GESTS ACTIXOAIETKA.

com - - s to sive a leaf-like appearance to the rays of the non-pigmented interradial
- " on the ventral surface of the eentrodorsal plate.

In another specimen of the type, however PI. VI. fig. 3\ the sides of the narrow
hasal gi res are almost parallel, and in every case meet at some distance within the
margin of the plate, while the interradial ridges resulting from their union scarcely
extend at all beyond the angles of the external pentagon.

-:ly, in the monstrosity represented in PI. VI. fig. 10, the dorsal aspect of which
is -een in PI. II. fie. S. both ridses and grooves are extremelv indistinct, and in no case
reach the outer margin of the plate ; while the margin of the internal cavity is markedly
pentagonal in form, and not ten-sided nor circular, as is the case where there are five
distinct interradial elevations alternating with the five radial areas (PI. VI. figs. 3, 8, 15) ;
and it does not project inwards so far as to conceal all the openings («) of the canals
leading to the marginal cirrhus-sockets. as is the case in the other two specimens of the
type (figs. 3j S and in var. 1 (fig. 15).

In all these three specimens of the type the ventral surface of the eentrodorsal plate
is not nearlv so flattened as in var. 1, but rises verv distinctly between its external and
its internal margins, while the radial areas are marked in the same wav bv various
indistinctly marked radiating ridges and furrows ; though as the floor of the central
cavity is also somewhat thicker, its depth is but little if at all greater. The same is the

se in the other three varieties, in each of which the basal grooves differ slishtlv in
form from one another and from the type. In every case they are widest about the
middle of their length, as in one of the specimens of the type (PL VI. fig. 8, b.g) ; this
is most marked in var. 3 (fig. 21), and least in var. 2 (fig. 17). They reach almost, if not
quite, to the margin of the plate, though the ridges formed by the union of their sides
extend but little if at all beyond it, except in var. 3 (figs. 20 & 21), in which two of the
angles of the external pentagon are marked by traces of the small processes (t) so
distinctly seen in var. 1 (figs. 11. 15). In this variety the course of the interradial
ridges is indistinctly visible on the dorsal surface of the plate (fig. 20), which is slightly
hollowed in the centre. The floor of the central cavity is, however, very thick and solid,
and its middle portion presents no trace whatever of any perforations for the canals of
~ing cirrhi, though those proceeding to the three marginal cirrhus-sockets are just
jle under the projecting lip (fig. 21, u), which conceals several others. These indicate
that more cirrhi either have been or would have been developed had the animal lived

_ . their external openings having been obliterated in the former case (the more
probable one) and not yet formed in the latter.

"• In Act. Solaris (PI. V. fie. 2) the ventral surface of the plate-like eentro-
dorsal piece is very nearly flat, as in. Ant. rosacea (PI. IV. fig. 15) and in Act. polymorpha,
var. 1 (PI. VI. fie. 15), rising but slightly from the circumference towards the centre,
and marked by five interradial elevations, along the top of each of which runs a long and
narrow basal gi Its width is tolerably uniform from its base until near its

end, where its sides suddenly approach one another, and meet at a little distance within
the margin of the plate, where the ridge formed by their union also ceases without
extending outwards beyond the general surface of the plate. The same is the case in

HE. P. H. CAEPEXTEE ON THE GENTS ACTEXOMETEA. 67

the large Act. robusta (PL Y. fig. 14), in which three of the hasal furrows are widest at
their central ends, and consequently triangular, while the others are somewhat irregular
in shape.

In both these specimens numerous small openings are visible on the floor of the
central cavity of the centrodorsal piece, but the principal ones leading to the marginal
cirrhus-sockets are concealed under its projecting lip. In the small centrodorsal piece of
Act. pectinate/., however, these last are very distinct (PI. Y. fig. 7, ?() and correspond
in number to the eleven external cirrhus-sockets (PI. Y. fig. 6, T7), so that the five
principal cirrhus-vessels leaving the quinquelocular organ would seem to divide at once
within the cavity of the centrodorsal piece, and not within the substance of its walls, as
is the case in Act. polynwrplta. The ventral surface of the centrodorsal plate of Act.
pectinata (PL Y. fig. 7) is by no means so flattened as in the closely allied Act. Solaris
(fig. 2), but rises considerably from the circumference towards the centre, and the inter-
radial ridges are well marked. The basal grooves {b.g) are narrow and parallel-sided, and
terminate within the margins of the plate, beyond which the interradial ridges are not
continued, so that there are none of the small processes extending outwards from the
angles as in some forms of Act. polymorpha. The median line of each of the radial areas
is occupied by a deep depression, which is particularly distinct at its central end. A
similar depression, though developed to a less extent, exists also in Act. polymorpha,Yax. '2
(PL VI. fig. 17, r.ar).

(§ 46) Xearly all the observers who have studied Comatula have regarded the " Knopf,"
or centrodorsal piece, as of essentially the same nature as the stem of the stalked Crinoids.
The first author who put forward this opinion was Schweigger1; and Miller's views2, pub-
lished two years later, were fundamentally the same, though somewhat modified in
form ; for the centrodorsal piece was regarded by Miller as composed of two separate
pieces, one forming the floor of the cavity and the other its sides and rim. The former
was described by him as a pentagonal imperforated plate, " analogous in situation to the
first columnar joint of the Crinoidea ; but as it is not required to transmit the passage
to the alimentary canal s (no prolongation of the column existing in this animal), it is
without central perforation."

The other or ventral half of the centrodorsal piece was regarded by Miller as an
annular or basin-shaped plate, representing the " pelvis" or basal circlet of JPentaerinus,
though he described it as marked externally by numerous sockets for the attachment of
the cirrhi, which in Pentacrinus are borne by the stem-segments only, and never by
the basals.

Goldfuss, who in most points followed Miller's views, differed from him considerably
with regard to the nature of the centrodorsal piece of Comatula mediterranean, which
they had both studied ; and his conclusions, though not absolutely correct, are much
nearer the truth than those of Miller. Pinding most specimens to bear three rows of

1 Op. tit. p. 64. - Op. tit. pp. 129,

:i It must be remembered that the canal which occupies the centre of the Crinoidal stem was originally sup; - . I
to be a continuation of the alimentary canal, and not, as we now know it to be, of the general perivisceral cavity
or coelom.

SECOND SERIES. — ZOOLOGY, VOL. II. 10

68 ME. P. H. CAEPENTEE ON THE GENES ACTINOMETEA.

dorsal cirrki, he described this species as having a column of three segments, and gave
a sectional figure in support of his statements L, which shows three segments below the
circlet of first radials, each bearing a row of cirrhi. It is doubtful how far this figure can
be relied on as accurate, though I have occasionally met with somewhat similar appear-
ances myself. Goldfuss, like Miller, was unacquainted with the remarkable condition
of the basals in this type; and as the "pelvis" described by Miller in Comatula was
rightly regarded by him as representing apart of the stem of Penlacrinus, he was led
to believe in the absence of basals in Com. mediterranea, though he found them in the
Com. multiradiata [Comaster), in which he described the rudimentary column as con-
sisting of only a single segment. Miiller was led, by his comparison of the component
pieces of the calyx of Comatula with those of the calyx of Pentacrinus aster/a (Captd-
medasce), to recognize the very close general correspondence between tbem ; and he
pointed out 2 that the presence of cirrhi at the upper end of the stem of the Penta-
crinoid larva on the one hand, and on the centrodorsal plate bearing the first radials
of the young Comatula on the other, indicate that the latter is comparable to the stem
of PentacriuHs, which bears the cirrhi in verticils separated by longer or shorter in-
tervals. This view of Midler's was pretty generally recognized as the true one, and it
was adopted and greatly strengthened by Wyville Thomson and Dr. Carpenter, who
came to precisely the same conclusions upon developmental grounds. The former
defined it as representing a " coalesced scries of the nodal stem-joints in the stalked
Crinoids," namely, of those joints which bear whorls of cirrhi, so that " the centro-
dorsal plate with its dorsal cirrhi in Antedon is the homologuc of the stem with its
cirrhi in the stalked Crinoids." Ludwig3 also, while referring to the development of
the centrodorsal as the enlarged uppermost stem-segment, speaks of it as " ein zusarn-
mengedrangter, oberer Stengelabschnitt, in welchem das verkalkte Gewebe keine Sonde-
rung in untereinander^eleo'ene Glieder erfahren hat."

(§47) The first rudiment of the stem of the Pentacrinoid larva as described by
Wyville Thomson 4 consists of a series of delicate calcareous rings forming a curved
line, which passes backwards from beneath the centre of the lower ring of plates, the
embryonic basals. Within each of these is formed a hollow sheaf of parallel calcareous
rods, united together by short anastomosing lateral branches ; the upper one of these, on
which the lower edges of the basal plates rest, soon becomes considerably wider and
thicker than the rest. " Daring the earlier stages of the growth of the Pentacrinoid it
is simply a circular band of the ordinary calcified areolar tissue, enclosing a sheaf of the
peculiar fasciculated tissue of the stem, gradually enlarging, with a central aperture con-
tinuous with the bore of the tube-like stem-joints."

This ring is subsccjuently developed into the permanent centrodorsal piece ; but the rudi-
ments of the first dorsal cirrhi do not appear around its lower contour until very much
later. New rings are developed immediately beneath it, until there are fifteen or sixteen

1 Petref. German, p. - " Lau des Pentacrinus, p. 10.

3 "Zur Anatomic ties BJiizo rinus lofot, nsis, "Zuitschr. fiir wiss. Zool. Bd. xxix. p. 127.

4 '• On the Embryogcny of Antedon rosaceus (Lin ck) (Comatula rosacea of Lamarck)," Phil. Trans. 1805, pp. 53G,
5o7.

ME. P. II. CAEPENTEE OX THE GENUS ACTINOMETEA. G9

ill all, the length and diameter of which are gradually increased by the deposition of new
calcareous material at their extremities and upon their outer cylindrical surface.

Dr. Carpenter 1 has shown that, at or about the period at which the suppression and
metamorphosis of the embryonic oral and basal plates begins, " the production of new
calcareous segments in the stem appears to cease, and a remarkable change begins to
show itself in the one on which the calyx rests. Instead of increasing in length, its
original annular disk augments in diameter, becoming convex on its lower surface and
concave on its upper, and it extends itself over the bottom of the calyx in such a manner
as to receive in its concavity the apices of the basal plates ;" and that portion of its
under surface which extends itself beyond the segments whereon it rests begins to be
marked by small tubercles, which are the origins of the dorsal cirrhi, while it also
" augments not only in absolute but in relative diameter, extending itself over the dorsal
or outer surface of the basal plates, which at the time of the detachment of the body

from the stem are almost entirely concealed by it A second whorl of cirrhi is now

developed, after the same manner as the first, between the latter (with which it alternates
in position) and the base of the calyx (pi. xlii. fig. 3), and a third whorl generally makes
its appearance before the detachment of the Pentacrinoid, so that the young Antedon
possesses ten cirrhi in different stages of advanced development, and from one to five still
rudimentary."

After the detachment of the young Antedon from its stem a minute five-rayed per-
foration is visible for a short time in the somewhat depressed central portion of the
inferior surface of the centrodorsal piece. It is the remains of the original " communi-
cation between the cavity of the basin-shaped plate and the central canal that is still
left in the upper segments (at least) of the stem. This perforation, however, is very
soon closed up by an extension of the calcareous network, so that no trace of it remains
visible either internally or externally."

We have thus seen that the centrodorsal piece " first presents itself in a form which
nowise differentiates it from the other joints of the cylindrical stem, but begins to take
on an extraordinary increase in a peripheral direction at the time when the dorsal cirrhi
first sprout forth, aud thenceforward remains in closer connexion with the calyx than
with the rest of the stem, from which it separates itself so soon as the dorsal cirrhi are
sufficiently developed to serve for the attachment of the animal." Each of these cirrhi
receives a " sarcodic thread, which proceeds from the sarcodic axis contained within the
cavity of the basin, and runs along the central canal of the cirrhus to its termination."

New cirrhi gradually appear between those previously formed and the base of the
calyx, and each receives a peduncle of sarcodic substance from the central axis ; and
" since the arrangement of the whole aggregate of such peduncles is distinctly verti-
cillate, the want of a definite plan in the grouping of the cirrhi on the external surface
of the centrodorsal plate seems attributable to their very close apposition."

During the whole period of the growth of the centrodorsal basin there is a " progressive
exuviation of the first-formed cirrhi from within outwards, concurrently with the develop-
ment of new ones near the margin, those cirrhi which surrounded the summit of the

1 Phil. Trans, he. clt. p. 732.

10*

70 ME, P. H. CAEPENTEE ON THE GENUS ACTINOMETEA.

stem being first shed and their sockets filled up by new deposit, and the space thus
formed being gradually widened by the progressive exuviation of the cirrhi that bound
it, and the filling up of their sockets." Thus the flattened central portion of the
dorsal surface of the plate by which it was originally attached to the joint of the stem
next beneath it increases very much in extent, and finally comes to bear a considerable
proportion to its diameter (PL IV. fig. 1). In Act. polymorpha (PI. VI. figs. 2, 7, 14,
16, 20), as we have seen, it extends over the whole of the dorsal surface of the plate, and
to a certain extent also in Act. robusta (PI. V. fig. 15). In Ant. Hschrichtii, however, it
does not reach any great extent, for most of the first-formed cirrhi do not appear to be
cast off as in Ant. rosacea and Ant. ecltica, or, if they are lost, their sockets are not obli-
terated, but they seem to be replaced by others, for I have frequently found young and
rudimentary cirrhi among the larger and perfectly developed ones around the central
portion of the large hemispherical " Knopf" of this species l.

(§ IS) In most pedunculate Crinoids, in which the calyx rests upon the uppermost
segment of the stem, this segment, instead of being the largest, is the smallest, being
the latest formed, while the base of the calyx is formed by the thickened and expanded
basals. Hence, as Dr. Carpenter remarks2, "it seems clear that the extraordinary
development of the highest segment of the stem into the centrodorsal basin, which is
characteristic of the mature Antedon, is connected with the multiplication of the pre-
hensile cirrhi which extend themselves from its dorsal surface."

At the base of the quinquelocular organ, and lying on the bottom of the centrodorsal
basin, but enclosed, together with the five chambers, in the above-mentioned fibrillar
envelope (JV), winch is probably of a nervous nature, there is, both in Antedon and in
Actinometra, a succession of verticils of five triangular leaflets 3. As already shown by
Dr. Carpenter, there can be little doubt but that the lower ones of these mark the
origins of the earlier cirrhal cords from the crinoidal axis. They increase in size from
below upwards, and from the extremities of some of the upper leaflets there issue groups
of three diverging cords that proceed to the cirrhi which are developed at a later period
around the periphery of the centrodorsal piece.

Greeff i has found the older cirrhus-cords still in connexion with these leaflets.
Apparently unaware of the original existence and subsequent removal of the cirrhi cor-
responding to them, he drew a distinction between the vessels which they enclose, and
which end close under the dorsal surface of the plate, and the vessels enclosed in the

1 I cannot altogether confirm Mailer's statement (' Gattung Comatula,' p. 239 (3) ) that the central apical portion of
the centrodorsal in Ant. Eschrichtii, where it was formerly united to the stem, may be covered with cirrhi. In all
the individuals of this species which I have examined (and they are many) there is always a small apical space
quite free from cirrhi ; it may not bo wider than the diameter of a large cirrhus-sockct, but it is always to be found.
I imagine that by the expression " Da es Antedon-Arten giebt, bei denen auch der mittlere Theil des C'entrodorsale
Cirrhen tragt (Antedon Eschrichtii v.. \\.) " (Crinoideen, p. 69, note), Ludwig does not mean any thing more than
that the centrodorsal is covered with cirrhi to a much greater extent than is usual in most Comatulce, where there is
generally a central space of considerable extent entirely free from cirrhi. Schliiter lias also expressed his doubts
respecting the accuracy of Midler's statement.

■ Proceedings 11. S., No. 100, 1870, p. 218. 3 These are seen in section in PI. VIII. figs. 3, 7.

4 Marburg Sifzungsberichte, No. 5, 1S76, p. 91.

ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA. 71

more peripherally placed cords proceeding from the upper leaflets, which enter the later-
developed cirrhi. Ludwig1 regarded them as rudimentary structures in Ant. rosacea,
because he found them in Ant. Eschrichtii to enter the more centrally placed cirrhi, which
are not removed, hut persist throughout life, as already mentioned. It will be evident,
however, from the facts stated above, that these cirrhus-cords, which end on the dorsal
surface of the plate, are not rudimentary structures, but the proximal ends of more
complete cords that have undergone a retrogressive metamorphosis, which in Ant.
Eschrichtii is not carried so far as in Ant. rosacea.

These facts all tend to strengthen the view first expressed by Wyville Thomson, that
the centrodorsal piece represents a coalesced series of the nodal stem-joints of the stalked
Crinoids. In Ant. Eschrichtii six or even more rows of cirrhus-sockets may be traced on
the hemispherical surface of the " Knopf," each row corresponding to a node in the
stem of Pentacrinus. Even in those Actmometrce in which only one row of sockets is
visible externally, the composite character of the centrodorsal piece is indicated by the
verticils of degenerate cirrhus- vessels at the base of the chambered organ (PI. VIII.
figs. 3, 7), and by the partially obliterated openings on the central part of the floor of
the centrodorsal cavity (PI. V. figs. 2, 7, II; PI. VI. figs. 3, S, 10, 15, 17, it).

Sometimes, indeed, the " Knopf " may actually assume a more or less columnar
form, as in the specimen of Ant. macrocnema mentioned in sect. II, and in the genus
Solanocrinus ; in both of which three or four alternating rows of cirrhi are visible.
In these forms we may reasonably suppose that the columnar centrodorsal was de-
veloped by the enlargement of the uppermost stem-segment on which alternating whorls
of cirrhi successively appeared, just as in Ant. celtica (PI. IV. figs. 1, 6, 8), Ant. rosacea
(fig. II), and Act. robusta (PL V. fig. 15), but not in such numbers as to obscure the
alternate arrangement (p. 69).

(§ 19) Gotte-, to whom we owe a series of most beautiful observations on the de-
velopment of the water-vascular system and perivisceral cavity of Comatula, has recently
questioned the accuracy of those observations of Wyville Thomson and Dr. Carpenter,
according to which the uppermost of the embryonic stem-segments developes into the
centrodorsal piece, and lias also attacked the view that it may possibly in some
cases arise from the fusion of two or more stem-segments as represented in Goldfuss's
figure.

His description of its origin is as follows : — " Die Anlagen der Centrodorsalplatte sind
schmale, aber doch netzformige Skeletstreifen, welche gleichzeitig mit den Basalia an
deren unteren Biindern entstehen und die obersten, noch eng zusammengedriingten
Stielgliederanlagen umgeben (fig. 13). Es ist daher spater, wenn diese Stelle sich
verschmachtigt, nicht immer ganz leicht, jene Anlagen der Centrodorsalplatte von den
obersten Stielgliedern zu unterscheiden. Beachtet man jedoch, dass sie anfangs das 5.-8.
Stielglied, und nachdem diese abwarts gcriickt sind, das 9., 10., 11., 12. u. s. w. umschlies-
sen, was Thomson iiberhaupt nicht erwahnt, so kann man sich der Ueberzeugung

1 Beitriige, loc. cit. p. 69.

2 " Vergleioh. Entwickelungsgesoli. d. Comatula mediterranca." Archiv, f. rnikrosk. Anat. Bd. xii. 1876, p. ij'.tl.

72 MR, P. H. CARPENTER ON THE GENUS ACTINOMETRA.

nicht verschliessen, das die Skeletzone, aus welclier die Centrodorsalplatte hervorgeht,
unabhangig von den eigentliclien Stielgliedern, mehr in Anschluss an die Basalia und
wohl als rudimentare Wiederholungen dersclben sich entwickelt. Besonders lehrreich
fiir diese Auifassung sind die gar nicht seltenen stiellosen Misshildungen der Comatula-
larven welche ich beobachtet babe. An solchen finden sich in der hinteren Korpcrhalfte,
welcbe ihre urspriinglichen Dimensionen behalt, statt der Stielglieder grosse netzformige
Platten welche den B,auin zwischen den Basalia und dem Endknopf ausfiillen (pi. xxviii.
fig. 50). Vergleicht man sie niit den viel schwachcren Anlagen der Centrodorsalplatte,
so spricht die Darstellung sehr an, dass sie durch die Stielbildung in ihrer Entwickelung
gehemmt und im umgekehrten Falle gefordert werdeu."

The only normal figure given by Gotte in support of his views represents a ciliated
larva, very much younger than the pentacrinoid stage, and with only eight stem-seg-
ments, over parts of the four uppermost of which are traces of a calcareous network
connected with the lower end of one of the embryonic basal plates. This network, which
reaches a more extensive development in the malformation represented in Gotte's other
figure, does not appear in any one of Wyville Thomson's figures of Coma tula larva?,
either in the free-swirnming or in the pentacrinoid condition. As his observations
were carried on for four years, in each of which he followed out the development of
several broods of embryos, it is impossible to suppose that he can have overlooked it had
it been present in the larva? of the British variety investigated by him. It is possible
that the early-formed irregular calcareous ring, " considerably wider and broader than the
ordinary rings of the stem, which lies immediately beneath the basal plates, and subse-
quently develops into the permanent centrodorsal plate," may represent the network
figured and described by Gotte. But then, as the latter says, Thomson makes no mention
of its extending downwards around the other stem-segments ; he gives, however, a series
of figures which, taken in connexion Avith those of the later stage given by Dr. Carpenter,
demonstrate conclusively that the above-mentioned ring does develope into the per-
manent centrodorsal piece. Gotte gives no figures whatever of the pentacrinoid stage.
If, as I believe to be the case, the network described by him as the rudiment of the
centrodorsal piece really does represent the primitive centrodorsal ring of Wyville
Thomson, commencing, be it remembered, as a network of small curved hollow spicules,
then his observations are in complete accordance with the views of "Wyville Thomson and
Dr. Carpenter. Gotte offers no explanation of its downward extension over the remain-
ing stem-segments as described by him in the Mediterranean variety ; and nothing of the
kind is described by the two above-mentioned observers as occurring in the British
variety, unless, indeed, it be the deposit of calcareous material upon the outer cylindrical
surface of each stem-segment by which its diameter is increased.

It is possible that this deposit might commence to be formed at an earlier period in the
Mediterranean variety than in the British one ; but it is difficult to understand its down-
ward extension from the rudiment of the centrodorsal plate as described by Gotte.

(§ 50) The condition of the centrodorsal piece in Ant. rosacea and in Acthwmelra
gives us, I believe, the means of understanding a problematical Cretaceous fossil, first

ME, P. H. CAEPENTER ON THE GENUS ACTINOMETRA. 73

described by Goldfuss \ of which neither be nor any subsequent observer bas given a
satisfactory explanation 2.

Glenotremites was at first placed by Goldfuss among the Echinoidea, and was sup-
posed by him to have some relationship with the Cidaridcs. It is a somewhat hemi-
spherical body, in the centre of the flattened upper surface of which is a large round
opening, called by Goldfuss the mouth. " Um den Mund liegen fiinf grosse ovale Locher
and zwischen diesen fiinf fiaehe Einnen, die sich bis zum llande erstrecken, wo ihre

Vertiefung nicht auslauft, sondern durch einen erhabenen Saum begriinzt wird

Die Locher gehen trichterformig in die Tiefe ; die Einnen sind die Felder der Fiihler-
gange." These grooves were supposed by Goldfuss to be perforated by minute pores for
the passage of tentacles.

The convex dorsal side of the body bears numerous sockets for the attachment of
cirrhi ; but Goldfuss compared these at first to the large tubercles of the Cidaridte. At
the apex are five smaller apertures ; and Goldfuss suggested that these might be respi-
ratory and the others genital, or, more probably, that both, like the cirrhus-sockets,
marked the points of attachment of various kinds of spines. Subsequently, however, in
his description of a second species, G. conoideus, he spoke of the larger apertures as
ovarian openings, and recognized the resemblance between the sockets on the convex
surface and the similar ones on the dorsal surface of the centrodorsal piece of Comatula
to which the cirrhi are articulated ; and he suggested that Glenotremites might be more
nearly related to the Comatulidee than to the JEchinidce, as he had at first thought.
Agassiz 3 adopted this view, and placed Glenotremites among the Crinoids, and near to
Comatula. Like Goldfuss, he regarded the central aperture as a mouth; but the five
punctated grooves radiating from it, which were supposed by Goldfuss to be provided
with tentacles, were regarded by Agassiz as the points of insertion of the radii. He did
not attempt to explain the five large openings on the ventral surface and the five smaller
apical ones. Reenter4, who, like all subsequent writers, accepted the view that
Glenotremites is the centrodorsal piece of a Crinoid allied to Comatula, regarded the
former as " trichterformigen Arm-Anfangen oder Mund-Winkeln," but did not under-
stand those of the dorsal surface.

D'Orbigny 5, who confused Glenotremites with Comaster and Solanocrinus under one
name, Comatula, and Pictet6, who retained it as a separate genus, did not attempt to
offer any further explanation of its peculiarities, and, so far as I know, Agassiz and
R center's views have been generally accepted.

1 Petref. German, i. p. 151), Taf. xlix. fig. 9, Taf. li. fig. 1, and ii. p. ISO, Taf. clx. fig. 18.

2 The following section was written early in 1877, and was in the hands of the Secretary of the Linnean Society
in June of that year. The substance of a portion of it was referred to in my paper on Pentacrintts and Rhizocrirms
('Journal of Anatomy and Physiology,' Oct. 1S77, p. 45). I am therefore exceedingly glad to find, from a paper
published early in 1878 (" TJeber einigc astylide Crinoiden," Zeitschrift der deutsehen geologischeu Gesellschaf t, Jahr-
gang 187S, p. 33), that Schluter has independently given the same explanation of Glenotremites as had occurred to
myself. I learn from his paper that even as late as 1871 Goldfuss's original views were still held by Geinitz (Elb-
thalgebirge, i. 1871, p. 91).

3 Prodrome, Joe. cit. p. 289. 4 Lethaja Geognostica, v. p. 177.

6 Cours elementaire, ii. p. 138. c Traite de Pale'ontologie, iv. p. 290.

71 ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA.

That Glenotremites is the centrodorsal piece of a Com a tula there can, I think, he little
donht ; hnt I see no reason to regard the central opening as a mouth, any more than in
any other of the centrodorsal pieces represented in Plates IV., V., and VI. In all these
cases the centrodorsal cavity, as we know from Gotte's observations, is derived from the
posterior part of the right peritoneal diverticulum of the larval alimentary canal, and
is therefore a part of the general body-cavity or enteroccel. It is occupied by the dorsal
half of the quinquelocular organ which rises through its central opening, the so-called
mouth of Glenotremites, and is continued as the "axial prolongation" (PI. VIII.
fig. 3, a.p) through the central aperture of the rosette upwards into the middle of the
visceral mass.

In Ant. rosacea this central opening is surrounded by five depressions (PL IV. fig. 15, q),
which are the dorsal terminations of the five radial diverticula of the body-cavity en-
closed between the radial spout-like processes of the rosette and the internal faces of the
first radials. These diverticula exist both in Antedon and in Actinometra (PL VIII.
fig. 3, a.r.c), but do not always reach the ventral surface of the centrodorsal piece as in
Ant. rosacea (PL IV. fig. 15). If we suppose the above-mentioned depressions (5) placed
radially around the centre of this surface to be deepened sufficiently to become openings
leading into the centrodorsal cavity, they would occupy precisely the same position as
the so-called genital openings J of Glenotremites ; and simply effect a more open com-
munication between the two parts of the coelom contained in the centrodorsal piece on
the one hand, and the general cavity of the calyx on the other, than when the ventral
surface of the former presents only a single central opening, as in Antedon and Acti-
nometra.

If the view advanced above be correct, it follows that the peripheral part of the areas
around the " genital openings " of Glenotremites are the representatives of the radial

1 These so-called " genital openings " were described by Goldfuss as " Locher." Schliiter, however, merely speaks
of them as " Gruben " (pp. 33, 42), and uses the same term for the whole cavity of the centrodorsal piece, " welche das
Herz (??) oder gekammerte Organ aufnimmt," and is therefore spoken of by him as the " Herzgrube." But from
his expressions, " centralo Herzgrube fiinfseitig " or " zehnseitig," he obviously intends "Herzgrube" to mean nothing
more than the central opening of the ventral surface of the centrodorsal, which he elsewhere calls the " Nahrungs-
canal " (!), although he evidently understands its real meaning.

I cannot therefore clearly make out from Schluter's paper whether the " Eadialgruben " are real perforations or
mere depressions, as in Ant. rosacea, which, by-the-bye, is the same species as the Antedon europceus of Greeff, and
not different from it as Schliiter seems to think. His figures (pi. i. figs. 1, 4, & 10, and pi. ii. figs. 1 & 3) appear to
represent ventral openings in the centrodorsal of some fossil Anteclons, just as described by Goldfuss in Glenotremites ;
but then he refers (p. 33) to Ludwig and Greeif (!) as describing the radial pits of Ant. rosacea as " sackforrnige, in
den Kalkseheitel eindringende blindgeschhssene Erweiterungcn der Leibeshohle."

His use of the word " blindgcschlossene " would seem to indicate that the " Eadialgruben " of his specimens are
really pits, closed below as in Ant. rosacea, and not actual openings; for in the latter case these extensions of the
ccelom contained within the radial axial canals would have opened into the centrodorsal cavity (also a part of the
coelom), i.e. into that part of it which was not filled up by the chambered organ, and they could not then be accu-
rately described as " blind.'' The " Eadialgruben " seem, however, to have been actual perforations in Ant. semi-
ghoosus (Sehl. pi. i. fig. 10); for Schliiter speaks of them (p. 42) as " mit der Centralgrube verschmolzen (reiehen
aber tiefer hinab)," though he suggests the possibility of this being due to an accidental fracture of their central bony
border. The point is one of some interest ; for in no recent Comatula yet known are the "Eadialgruben" more than
simple pits, such as arc generally found in Ant. rosacea.

ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA. 75

areas on the ventral surface of the centrodorsal piece of Ant. rosacea (PL IV. fig. 15, r.ar),
in each of which lies the convex dorsal surface of a single first radial.

What, then, are the five radiating punctated grooves which Agassiz and Poemer
regarded as the articular surfaces for the attachment of the five arms of Glenotremites?
I befieve them to be the representatives of the basal grooves on the ventral surface of
the centrodorsal piece of Actinometra (PL V. figs. 2, 7, 11, PL VI. figs. 3, 8, 10, 15, 17,
21, b.g). They are sometimes slightly developed in Ant. rosacea, one lying between every
two of the depressions (q) mentioned above (PL IV. fig. 15), in precisely the same manner
as the grooves and " genital openings " alternate on the ventral surface of Glenotremites.

We thus see that the peculiarities of the ventral surface of Glenotremites may be
readily explained by what we know of the corresponding parts in Antedon and Actino-
metra. The apertures in the centre of the dorsal surface admit of an equally simple
explanation.

The quinquelocular organ forming the dorsal termination of the axial prolonga-
tion of the adult Qomatula consists of five chambers, arranged around a central axis
which contains numerous vessels. In Pentacrin/is there is no centrodorsal piece, but the
quinquelocular organ is contained in a cavity, the sides of which are formed by the first
radials above and by the basals below. Its five chambers are not closed below, but
narrow considerably, and are continued down the stem as five long vessels arranged
symmetrically around a central axis. The same appears to be the case in the stem of
the Pentacrinoid larva of Qomatula ; for, as already mentioned, Dr. Carpenter has
described a minute five-rayed perforation occupying the central portion of the dorsal
surface of the recently detached Antedon. I regard this perforation as homologous with
the five small apertures arranged around a single central one on the dorsal surface of
the centrodorsal piece of Glenotremites, and with the similar openings on the underside
of the calyx of the other stalked Crinoids — for example, of Ciqrressocrinus. The fact
that the young Antedon rosacea has only three rows of dorsal cirrhi when liberated
from its stem, while there are four or six rows on the dorsal surface of Glenotremites,
does not at all tell against this view. Indeed Sars1 has shown that the pentacrinoid stage
persists in Antedon Sarsii very much longer than in Ant. rosacea, and he has found
specimens with nearly thirty cirrhi still in a pedunculate condition, the cirrhi being
placed in such close proximity to one another that any trace of a distinct order in their
arrangement was entirely obliterated. The exterior of the centrodorsal piece of Ant.
Sarsii, therefore, immediately after its liberation from the stem, would present (its size,
of course, excepted) a very similar appearance to the convex dorsal surface of Gleno-
tremites, viz. a central five-rayed opening, or possibly even a single opening with five
others round it, the rest of the surface being covered with sockets for the articulation of
the dorsal cirrhi.

(iv.) The Pentagonal Base of the Calyx.

(§ 51) In all the Actinometra? with which I am acquainted the Pentagonal Base of the
calyx formed by the close mutual adhesion of the five first radials, together with the

1 ' Criuoides vivants,' p. 57.
SECOND SERIES. — ZOOLOGY, VOL. II. 11

76 ME. P. H. CAEPENTEE ON THE GEM US ACTINOMETEA.

rosette or metamorphosed basals, differs in many points from that of Ant. rosacea and
of all the other species of Antedon which I have examined.

In all of these the external or distal faces of the first radials slope at a considerahle
angle from above and within downwards and outwards, so that a view of the upper or
ventral aspect of the radial pentagon formed by their union (PI. IV. figs. 4, 17) shows,
not only their small superior or ventral faces around the central funnel-shaped space (F),
but also the greater part of their inclined external faces (PI. IV. figs. 6, 8, 14), viz. the
fossae (/) for the attachment of the muscles between the first and second radials (PI. IV.
fig. 5, r.m) and the smaller ones (/*) which lodge the interarticular ligaments, the distal
opening of the central canal (c.c), and the large transverse articular ridge (i), together
with more or less of the large fossa (/) which lodges the elastic ligament. The amount
of this fossa which is visible on the ventral aspect of the radial pentagon varies in
different cases.

In correspondence with this inclination of the distal faces of the first radials of Antedon
to the vertical or dorsoventral axis of the calyx, their ventral faces are much reduced and
are very small in comparison with the dorsal ones.

When the ventral surface of an isolated first radial of Ant. rosacea is examined (PI. IV.
fig. 12 a), it is seen to be divided into a central and a peripheral portion by two curved
ridges, bending towards each other along the median line, and there separated by a
furrow (fj). The central portion only is the true ventral face of the radial. It slopes
inwards, so as to contribute to the formation of the central funnel-shaped space (PI. IV.
fig. 17, F) occupying the centre of the radial pentagon, and partially filled up by a cal-
careous network formed by the inosculation of processes which proceed from the internal
and ventral faces of the surrounding radials (PI. IV. fig. 12 a, en). The peripheral
portion, on the other hand, slopes outwards, and is, in fact, the upper or ventral half of
the inclined external face (fig. 14), namely, the large vertical lamella? in which the mus-
cular fossa? are excavated. The upper and inner edges of these lamella? form the curved
ridges above mentioned, each of which has two limbs, one superior (fig. 12 a, g^), which is
horizontal, or nearly so, and forms the external boundary of the ventral face, and one
inferior (g2), which descends along the median line of the inclined external face towards
the great transverse articular ridge (i) ; it is separated from the corresponding inner
edge of the other muscular fossa by the furrow (/J, which may therefore be called
" intermuscular."

These curved ridges thus produce great inequalities in the ventral aspect of the
radial pentagon (PL IV. fig. 17). The walls of its central funnel present an alternation
of radiating ridges and furrows, of each of which there are ten. Five of the furrows
(v.i.f) correspond Avith the divisions between the component pieces, and are therefore
interradial, while the ridges which bound them are the superior limbs (#,) of the curved
ridges above mentioned, one belonging to each of the two contiguous radials. Of the
other five furrows, one passes along the middle of each of the five radials, and both the
ridges which bound it belong to the same piece — being, in fact, the median or descending
margins (g2) of the large vertical lamella? in which the muscular fossa? (f) are excavated.
These furrows, therefore, are simply the intermuscular furrows of the distal faces (PL IV.

MR. P. H. CARPENTER OX THE GENUS ACTINOMETRA. 77

figs. 12 a, 14, fj), and they only appear on the ventral aspect of the radial pentagon
because of the inclination of these faces to the vertical axis of the calyx. They do not
appear to exist in Ant. celtica (PI. IV. figs. 4, 6, 8), in which the radial muscles are
far larger than in Ant. rosacea, so that the vertical lamellae to which they are attached
attain a ninch greater size. These are placed at such an angle to the dorsal portions of the
distal faces of the ra dials that they stand up nearly vertically, and form the outer wall
of the central funnel-shaped space (fig. 4, F) which leads downwards into the cavity of
the centrodorsal piece. Its pentagonal rim is formed, as in Ant. rosacea (fig. 17), hy
their superior margins, two of which, belonging to contiguous radials, bound each of the
five ventral interradial furrows (v.i.f) which mark the angles of the pentagon. In the
centre of each of the five sides of this pentagonal rim is a deep notch (fig. 6, f2) ; but it
does not descend on to the distal face of the corresponding radial so as to form an inter-
muscular furrow bounded by the median descending margins of the muscular fossa?, as
in Ant. rosacea (figs. 14, 17, /I) ; for these fossae are so large, and extend so far towards
the median line, that their inner margins unite and form a prominent vertical ridge
(figs. 4, 6, //;;), which passes below into the elevated rim around the opening of the central
canal (c.c).

(§ 52) These five notches in the sides of the pentagonal rim of the central funnel of
the radial pentagon in Ant. celtica (PI. IV. fig. 6, f2) represent the points at which in
Ant. rosacea the superior or central end of each intermuscular furrow (f_) passes at a
slight angle, due to the inclination of the distal face, into a shallow depression (figs. 12 a,
17, v-r.f) occupying the centre of the small ventral face of each first radial. This depression,
which is much better developed in Actinometra, is far more distinct in some specimens of
Ant. rosacea and Ant. celtica than in others, and in the dry state is barely visible.
When, however, the interior of the calyx is viewed from above after the visceral mass
has been removed, so as to lay open the circumvisceral ccelom, and expose the ventral
aspect of the radial pentagon, the position of the ventral radial and interradial furrows is
indicated by dark lines converging towards the centre (fig. 5). These are due to the fact
that the parietal layer of the peritoneum which lines the interior of the calyx descends
into these depressions, so that its pigment is here more thickly aggregated than on the
rest of the ventral surface. A similar slight depression lined by the pigmented peri-
toneum exists on the median line of the ventral face of the second and third radials and
of the basal brachial segments, and it lodges the dorsal portion of the cceliac canal, which,
in the intervals between the segments, sends down diverticula between the muscles
connecting them, so that its course is readily traceable by the greater intensity of the
pigment along the median bine of the segments and between the two muscles connecting
every pair (fig. 5, v.r.f). At the base of the arms the cceliac canal becomes broken up
by connective-tissue septa into a number of intercommunicating spaces, which open
freely into the general cavity of the calyx or circumvisceral ccelom. The dorsal part of
the canal, however, retains its primitive relation to the skeleton and muscles, and is
lodged in the furrows on the ventral faces of the radials (fig. 5).

We have already seen that the inner wall of the funnel-shaped space (F) occupying the
centre of the radial pentagon is formed by the inclined ventral faces of the five first

11*

78 MR. P. H. CARPEXTER OX THE GEXES ACTIXOMETRA.

radials (figs. 4, 12 a, 17). These are not simply plane, but are usually more or less
divided up by delicate calcareous processes which extend to meet the ventral face of the
rosette, and collectively form a complicated network (c.n), filling up the central funnel,
and often partially bridging over the ventral radial furrow, so as to convert it into an
incomplete canal.

At the inner margin of the ventral face this furrow turns downwards, and passes
directly into a nearly vertical furrow occupying the median line of the proximal or
internal face (PL IV. fig. 12 c, a.r.f), and more or less completely converted into a canal
by the union of irregular processes, which extend themselves from its sides to meet the
rosette. As it descends towards the dorsal face and passes between the inner raised edges
of the two apertures {/, y) of the central canal, this axial radial farrow becomes a complete
canal, for its edges are closely applied to the inflected margins of. one of the five radial
spout-like processes of the rosette (PI. IV. figs. 13, 16, p).

The five canals thus formed may hence be regarded as enclosing cavities directly con-
tinuous with the cceliac canals of the arms, in the direction of which they lie ; and they
thus enclose portions of the body-cavity, which I will call the radial ccelom1.

They open on the dorsal surface of the radial pentagon by five large openings (PI. IV.
fig. 16, Q), that correspond with five more or less distinctly marked circular depressions,
which are placed radially on the ventral surface of the centrodorsal piece around the
margins of its central cavity (fig. 15, q), and the cauals end blindly in these depressions.
"Where these canals are enclosed by the spout -like processes of the rosette, they are com-
pletely shut off both from one another and from the dorsal extension of the ccelom which
occupies the central funnel-shaped space within the radial pentagon (figs. 1, 17, F), and
passes down into the cavity of the centrodorsal piece through the central opening of the
rosette (fig. 16, r.o). On the ventral side of the rosette, however, these radial axial canals
are only partially complete, and are in free communication with the numerous plexiform
spaces into which the funnel-shaped space is broken up by the above-mentioned cal-
careous network. The central portion of this system is very irregular; but peripherally
the plexus becomes more regular, and five axial interradial canals are traceable between
the five radial ones, with which, as with the centre of the plexus, they are in free
communication.

These interradial canals are continuous with the interradial furrows which are visible
on the ventral aspect of the radial pentagon (PL IV. figs. 4, 17, i'-i-f), and they enclose
diverticula of the circumvisceral ccelom to which the name interradial ccelom may be
given. They do not descend so far towards the dorsal surface as the axial radial canals,
and are not, like the latter, enclosed (normally, at any rate) by spout-like processes of
the rosette ; for their course towards the dorsal surface is terminated by the five short
triangular processes of the rosette (figs. 3, 7, 13, 16, o), which are directed towards the
sutures between the five radials.

(§ 53) This is well seen in Ludwig's schematic vertical section through the body of
Ant. rosacea-, in which the radial ccelom (Lr) is rightly represented as both longer and

1 The general relation of these axial radial canals is precisely the same in Act inometra as in Antedon. See PI. VIII.
figs. 3, 6, a.r.c. 2 Beitrage &c. Taf. xix. fig. 74.

ME. P. H. CAEPEXTEE ON THE GENITS ACTEXOMETEA. 79

larger than the interradial coelom (Id). It is also seen in Taf. xv. figs. 25, 26, on a larger
scale ; and in Taf. xiv. figs. 20-24 hoth the radial and the interradial diverticula of the
body-cavity are seen in transverse section, separate from one another towards the dorsal
side, but communicating freely nearer the ventral surface, both with one another and
with the centre of the plexus. In figs. 20-24, Ludwig has accidentally lettered them
L' and L" respectively. This is unfortunate, as these letters are employed by him in
his other figures to designate the circumvisceral and axial body-cavity ; while in fig. 26
he uses the same letter L to designate the system of plexiform spaces occupying the
central funnel of the pentagonal base, as he employs in his other figures for the inter-
visceral division of the body-cavity.

This hardly agrees with his text ; for on p. 43 he says : — " Ueber den ersten Eadialien
lost sich die axiale Leibeshohle in eine Sumnie von mit einander allseitig communiciren-
den Maschen raumen auf, welche zwischen die ersten Eadialien eindringen, bier das
dorsale Organ [i. e. axial prolongation] umgeben und endlich mit zehn blmdgeschlos-
senen Fortsetzungen endigen, von denen fiinf radiar gerichtet sind (Lr), funf interradiar
(Id). Der Dorsalcanal [ = cceliac canal] des Amies giebt seine Lage dicht fiber den
Kalkgliedern und zwischen und fiber deren Aluskelpaaren nicht auf bis er fiber dem
ersten Eadiale angekommen ist, wo er sich gleichfalls in die schon erwahnten Maschen-
raume auflost. Letztere stehen also in Verbindung mit der axialen Leibeshohle und mit
den Dorsalcanalen der Arine, aber sie dehnen sich auch ferner nach oben und seitlich aus,
und erfiillen bier den Raum der rings um die axiale Leibeshohle zwischen dem Yentral-
canal und dem Dorsalcanal in der radiaren Halfte, zwischen Ventralperistom und Dorsal-
peristom in der interradiaren Halfte der Scheibe fibrig bleibt." This space, the general
perivisceral cavity, falls naturally, as Ludwig has pointed out, into two divisions — one
external or circumvisceral, between the visceral mass and the body-wall, and corresponding
to the " cceloni " of Dr. Carpenter ; and one internal or intervisceral, surrounding the axial
body-cavity (or axial canal of Dr. Carpenter), and occupying the spaces between the
various coils of the alimentary canal within the visceral mass. This last corresponds
to the intramural spaces and mesenteric sinuses of Dr. Carpenter ', and not to the former
only, as Ludwig appears to think (p. 55).

Of all the divisions of the body-cavity this intervisceral cceloni is the one which is
least directly connected with the plexiform network between the first radials (Ludwig,
figs. 26, 74, L) and with the cceliac canals of the arms ; for it is completely separated from
the latter by the visceral layer of the peritoneum, except at the minute aperture in the
under surface of the visceral mass, where the axial prolongation, coming up from the quin-
quelocular organ through the central vacuity of the pentagonal base, enters the inter-
visceral cceloni contained within the visceral mass. "When the latter is turned out of
the calyx the intervisceral ccelom contained within it is, of course, removed at the same
time, while the plexiform system of spaces between the first radials, and the continuations
of the cceliac canals of the arms which terminate in it, are laid open ; both of these, there-
fore, are manifestly portions of the general c/;'c^//iviseeral cavity surrounding the visceral
mass. Ludwig, however, makes the following statement (p. 90) : — " Die Hauptab-

1 Proc. E. S. no. 166, 1S76, pp. 216, 217, 225.

80 MR. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

sclmitte der Leibeshohle in Scheibe und Arnien stelien niiteiuander paarweise in engerer
Bezielnrng, indem sick dieaxiale Leibeshole fortsetztin die Ventralcanale [subtentacular]
der Arnie und Pinnulse, die interviscerale in die Dorsalcanale [cceliac] und die eircum-
viscerale in die Genitalcanale."

I cannot corroborate this statement except with regard to the axial body-cavity, the
connexion of which with the subtentacular canals of the arms was first shown by
Dr. Carpenter.

The ventral portion of the circumvisceral body-cavity, viz. the limited and much
divided space between the parietal and visceral layers of the ventral peritoneum, certainly
does stand in direct connexion with the genital canals of the arms ; but its dorsal
portion, viz. the space between the visceral mass and the skeleton of the calyx, is, as
already mentioned, far more directly a continuation of the cceliac than of the genital
canals. The former gradually increase in size as they approach the disk, becoming very
large in the second and first arm-segments, and traversed by numerous connective-tissue
bands, which are directly continuous with those of the circumvisceral space ; while the-
genital canal remains relatively small, and is nothing more than a space in the horizontal
septum separating the subtentacular and cceliac canals.

The beautiful investigations of Gotte x have shown that the primitive ccelom of the
pentacrinoid larva of Antedon consists of two parts : (1) an oral or ventral one, de-
veloped from the left peritoneal diverticulum of the primitive alimentary canal ; and
(2) a dorsal one, which sends a prolongation backwards into the stem, and is developed
from the corresponding right peritoneal diverticulum. These divisions of the primitive
coelom had been previously described by Dr. Carpenter -, Metschnikoff33 and Greeff 4,
to all of whom, however, their origin was unknown. The last observer regarded the
ventral division as " den vom Wassergefasssystem und der hinteren Leibeshohle geschie-
denen urspriinglichen Blutsinus ;" for he supposed it to be continuous with the cavity
of the axial prolongation, which he called the " dorsovcntral Gefassstrang." Dr. Car-
penter has found, however, that this structure breaks up into five branches, one of which
goes to each of the primitive rays, and developes into the so-called " genital rachis " of
the arms, while the oral ccelom of the pentacrinoid larva (the "Bloodsinus " of Greeff),
sends off an extension into each of the arms as its subtentacular canal. In the direction
of the radii it forms, of course, the subtentacular canals of the disk ; but elsewhere, or
interradially, it becomes gradually limited by the enlargement of the visceral mass, and
by the formation of adhesions between its upper surface (visceral layer of the peritoneum)
and the parietal layer lining the under surface of the ventral perisoine ; so that the
ventral portion of the circumvisceral ccelom enclosed between these layers, to which the
primitive oral ccelom gives rise, becomes vrery much reduced in extent. We do not yet
know the precise origin of the genital canals of the arms ; but it seems most probable
that, like the ventral portion of the circumvisceral coelom with which they are connected
in the disk, they are developed out of the lower or dorsal half of the extension into the

1 Op. eit. p. 591, Taf. xxvi. fig. 19. 2 Phil. Trans, he. eit. p. 728 ; Proc. R. S. no. 166, p. 228.

3 " Beitr. z. Entwickelungsgesch. einiger niederen Thiere," Bull, de l'Acad.. Imp. des Sciences de St. Petersb. torn. sv.
1871, pp. 502-509. 4 Marburg Sitzungsbericbte, 1876, No. 5, p.

ME. P. H. CAEPENTEE OX THE GENUS ACTINOMETRA. 81

arms of tlie primitive oral coelom — the upper or ventral portion of which gives rise to
the suhtentacular canals of both arms and disk.

The dorsal or aboral coelom of the pentacrinoid larva lies beneath the annular
mesentery, and forms the dorsal half of the primitive circumvisceral coeloin, long before the
alimentary canal is sufficiently convoluted to give rise to a distinct intervisceral coelom.
Like the oral coelom it sends off radial extensions into the developing arms, but beneath
the horizontal partition extended from the mesenteric bands, and these become the cceliac
canals. Its dorsal prolongation gives rise to the cavity of the centrodorsal piece, and
ultimately to the central canals of the calcareous segments. Both of these, together with
the plexiform space between the first radials and the cceliac canals converging to it, are
therefore, like the greater part of the circumvisceral coelom, derived from the right peri-
toneal diverticulum of the primitive digestive cavity ; while the left one gives rise to the
suhtentacular canals of the disk and arms, and to the ventral portion of the circum-
visceral coelom. The primitive distinction between the oral and the aboral ccelom of the
larva, indicated by the mesenteric fold, becomes, however, gradually obliterated by the
development of numerous similar septa of connective tissue, and by the growth of the
alimentary canal and its consequent winding.

The axial canal, continuous above with the oral, and below with the aboral ccelom, is
produced by the limitation of the central space left by the coiling of the intestine around
the stomach ; while the remainder of the spaces between the coils become the inter-
visceral ccelom, which is therefore not developed to any extent until after the cceliac
canals of the rays have been extended from the primitive aboral coelom.

(§ 51) All the species of Antedon do not agree with Ant. rosacea and Ant. celtica in
the great inclination of the distal faces of the first radials to the vertical axis of the calyx,
so that these faces enter into the composition of the ventral aspect of the radial pentagon.
In a new and undescribed Antedon from the Philippines this inclination is very slight;
and in a view of the pentagonal base from above but little more is seen than the proper
ventral faces of the component radials. In this respect, therefore, forms such as these
present an approximation to Actinometra (PI. V. fig. 1, PI. VI. figs. 5, 12, 23), in which
the distal faces of the first radials are nearly or quite vertical, and do not at all enter
into the composition of the ventral aspect of the pentagonal base, which consists simply
and entirely of the five adjacent ventral faces of the component radials.

In Ant. rosacea (PL IV. fig. 17) these form a five-pointed star, the central surface of
which slopes rapidly downwards and inwards as the inner wall of the central funnel (F) ;
while its five rays correspond with the divisions between the component radials, and are
bounded by the large ridges which form the upper and outer margins of the two adjacent
muscular fossa? (/) of every pair of contiguous radials (PI. IV. fig. 12 a, gx). The sutures
between the radials are marked by slight depressions of their ventral surfaces, and these
are completed by the ridges at their sides into the ventral interradial furrows already
described (PI. IV. figs. 1, 17, v.i.f) ; they occupy the five rays of the star, and alternate
with the five shallow depressions (v.r.f) lying in the centre of the ventral faces of the
first radials. In Actinometra, as will be seen further on, these depressions become very
marked ; but in Ant. rosacea they are hardly deep enough to be called furrows, and are

82 ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

generally more or less concealed by the calcareous network occupying the opening of the
central funnel. They are the ventral continuations of the five canals enclosed by the
radial spout-like processes of the rosette, and they pass downwards and outwards in the
reentering angles of the star into the intermuscular furrows on the distal faces of the
component radial s (PI. IV. figs. 12 a, 1-1, 17, f). These reentering angles, which are
bounded by the superior margins of the two muscular fossse of each radial, are more open
in Ant. celtica (fig. 1) than in Ant rosacea (fig. 17), so that the rim of the central funnel
becomes more nearly pentagonal, having somewhat the shape of a Goniaster. This is
still more marked in Ant. Esckrichtii, while in Actinometra it becomes a regular pentagon
(PI. V. fig. 4, PI. VI. figs. 5, 12, 23).

In correspondence with the nearly vertical position of the distal faces of the radials in
Actinometra, their ventral faces, which in Ant. rosacea and Ant. celtica have a very
steep inward slope, occupy a nearly horizontal position, sloping but very gently inwards
towards the central space, so that the opening of the funnel becomes widely expanded.
Its inner walls, formed by the adjacent ventral faces of the contiguous radials, which are
relatively much larger than in. Antedon, are generally more or less sculptured out into
a series of radiating ridges and furrows, the number and distribution of which vary in
different species.

(§ 55) In Act. pectinata the ventral surface of each first radial (PI. V. fig. 9 a) is
nearly as even and regular as that of Ant. celtica (PL IV. fig. 4) or Ant. rosacea (PI. IV.
figs. 12 a, 17), and in some cases it may be even more so. It is, however, both absolutely
and relatively larger, as it is not encroached upon by the distal face, which stands nearly
at right angles to it, and the furrow (v.r.f) occupiug its median line is far more distinct
than in either of the two species of Antedon. These points are clearly seen in a com-
parison of figs. 5 on Plates IV. and V., which represent the ventral aspect of the calyx,
as seen after removal of the visceral mass, in Ant. celtica and Act. pectinata respectively.
In the former (PI. IV. fig. 5) the second and third radials and the bases of the arms are
at a higher level than the pentagon of the first radials, owing to the inclination of the
distal articular faces of the latter ; but in Actinometra (PI. V. fig. 5) the whole ventral
surface of the calyx is iu one horizontal plane, as the opposed articular faces of the first
and second radials are parallel to the vertical axis of the calyx, and not more or less
inclined to it, as in Antedon.

We have seen that in Ant. rosacea the lateral margins of the ventral faces of the first
radials (PL IV. figs. 12 a, 17) are somewhat depressed, so that when two pieces are in
contact a shallow interradial groove marks their line of union on the ventral side. It
is deepened into a furrow (v.i.f) by the elevation at its sides of the ridges forming the
superior margins of the muscular fosste of the inclined distal face. This interradial
depression also occurs in Act. pectinata (PL V. figs. 5, 9 a &c.) ; but as the vertical
lamella; (fig. 9 c, g) in which the muscular fossae are excavated are very small, and do
not extend inwards so as to encroach upon the ventral face, there are no ridges at the
sides of this interradial depression (fig. 5, v.i.f) converting it into a deep furrow as in
Ant. rosacea (PL IV. fig. 17), so that it is far less conspicuous than the corresponding
radial furrow (PL V. figs. 5, 9 a, v.r.f).

MR. P. H. CARPENTER OX THE GENUS ACTLXOMETRA. 83

The ventral aspect of the radial pentagon of Act. Solaris (PI. V. fig. 4) consists almost
entirely of the conjoint ventral faces of its component pieces ; the distal faces are very
slightly inclined to the vertical axis of the calyx, so that portions of the fossae lodging
the radial muscles and the interarticular ligaments hecome visihle (PL Y. fig. 4?,f, h).
The ridges (g{) which bound the muscular fossa? superiorly form by their apposition the
outer margin of the ventral surface of the pentagonal base, which is interrupted at ten
points, five being radial and five interradial. The former, which lie between the two
muscular lamellae of each radial, indicate the union of the intermuscular furrows of the
distal face with the ventral radial furrow occupying the median line of the superior face
(PL V. fig. 1, v.r.f) ; while the latter, which are at the angles of the pentagon, are the
outer ends of the ventral interradial furrows (v.i.f) corresponding with the sutures
between every two contiguous radials, the superolateral edges of which are slightly cut
away, so that by the apposition of every two pieces an interradial furrow is formed.

These interradial furrows, like the radial ones, slope gently inwards towards the centre.
The two sets, as soon as they pass into the axial furrows on the internal faces, become
respectively converted into five radial and five interradial axial canals by the union with
one another in successive pairs of small processes extending from the intervals between
them towards the central calcareous network (PL V. fig. 1, c.n). These processes are
the central ends of ridges which are developed on the two halves of the ventral surface of
each first radial, between its median furrow {v.r.f) and its lateral margins. The small
and irregular furrows between them usually converge towards the radial or interradial
furrows, where they begin to descend into the corresponding axial canals; but in two
cases (PL V. fig. 4, x ) they are also converted into canals by the small bridge-like pro-
cesses above mentioned. These intermediate canals, like the normal radial and inter-
radial ones, are in free communication with the rest of the spaces in the calcareous net-
work, just as in Ant. rosacea; but the radial ones do not extend so far towards the
dorsal surface of the pentagonal base as in this species, as will be seen when we come to
study its dorsal aspect (PL Y. fig. 3).

(§ 56) In Act. robusta (PL A*, fig. 11) this sculpturing or development of ridges and
furrows on the ventral faces of the first radials is carried much further than in Act.
Solaris. The muscular fossae (f) are also somewhat deeper, and the median radial
furrows which proceed inwards from the intervals between their superior margins (g^)
along the ventral faces of the radials are broken up very soon into a number of small
irregular furrows ; all converge, however, towards the centre, by the development of
numerous ridges of a similar nature to those rising from the lateral halves of the ventral
faces in Act. Solaris.

These ridges completely obliterate all traces of any regularity in the passage of the
radial furrows into the central calcareous network (en), as was so marked in Act. Solaris
(fig. 4). The interradial furrows, too, are not particularly distinct, as the ventral surfaces
of the radials fall away but little towards their lateral margins. Towards the centre,
however, they become more distinct, and arc bridged over by processes extended from
the above-mentioned ridges, so that they pass downwards as canals into the system of
plexiform spaces occupying the central funnel of the pentagonal base. The position of

SECOND SERIES. — ZOOLOGY, VOL. II. 12

84 MR. P. H. CAEPEXTEE OX THE GENUS ACTINOMETRA.

one of these axial interradial canals is indicated, in PI. V. fig-. 11, by a brown bristle (II)
which has been passed along it.

In the type of Act. polymorpha the distal faces of the radial pentagon are placed
somewhat more vertically than in Act. Solaris, so that scarcely any trace of the muscular
fossae is to be seen on its ventral aspect (PL VI. fig. 5). This is still more the case in
var. 4 (fig. 23), in which the ventral aspect of the radial pentagon exhibits nothing but
the extremely sculptured and inclined ventral faces of its component pieces ; it is divided
into a very large number of ridges and farrows, nearly every one of the latter having a
canalicular ojiening into the central network (c.n). The radial furrows arc thus entirely
obliterated; and as there is no corresponding intermuscular furrow on the distal face
(as in Ant. celtica, PI. IV. figs. 4, 6), there is nothing to indicate their position on the
outer margin of the radial pentagon. But the interradial furrows (PI. VI. fig. 12, v.i.f)
are readily distinguishable by their being somewhat deeper and straighter than the
secondary radial furrows. This is also the case, but to a less extent, in the type (fig. 5)
and in varieties 2 and 3 ; but var. 1 is somewhat different, and in this respect approaches
Antedon rosacea more than any other Actinometra with which I am acquainted. The
distal faces of the radial pentagon (PI. VI. fig. 12) are perceptibly inclined to the
vertical axis of the calyx, so that even the opening of the central canal (c.c) appears on
its ventral aspect. The muscular fossae (/) are deep, so that their superior margins
project inwards and encroach somewhat on the ventral faces ; and the median furrows
of the latter are tolerably deep, their outer extremities passing over into the inter-
muscular furrows (f) of the distal faces. The interradial furrows between the elevated
lateral halves of the ventral faces (PI. VI. fig. 12, v.i.f) are also deep, but the ventral
faces are plain and scarcely at all sculptured, so that both radial and interradial furrows
pass down with tolerable regularity into the peripheral axial canals of the central cal-
careous network (c.n).

(§ 57) In Antedon rosacea, as we have already seen, the five radial diverticula of the
ccelom terminate blindly on the ventral surface of the centrodorsal piece in five depres-
sions (PI. IV. fig. 15, q), which are disposed around the opening of its central cavity (ccl.c).
In correspondence wdth these depressions the dorsal surface of the pentagonal base pre-
sents five large openings (PI. IV. fig. 16, Q), disposed in like manner around the margins
of its central space. These openings are the dorsal terminations of the five radial axial
canals, and are formed by the application of the five radial spout-like processes of the
rosette (figs. 13, 16, p) to the inflected margins of the two openings (.(•', y) on the internal
face of each first radial (fig. 12 c), through which the secondary basal cords (X, 1\) pass
on their course from the fibrous mass enveloping the quinquelocular organ to the
circular commissure contained within the radial pentagon (compare PI. VIII. fig. 2).

The existence of these five large openings (PL IV. fig. 16, Q) is due to the fact that the
dorsal face of each first radial presents a deep notch in the centre of its inner margin
(fig. 12 b, Q') ; this notch indicates the continuation towards the dorsal surface of the
radial axial furrow on the internal face (fig. 12 c, a.r.f) ; and when this furrow becomes
converted into a canal by the application to its inflected edges of one of the spout-like
processes of the rosette, the notch on the dorsal face also becomes converted into a cir-

ME. P. H. CARPENTER OX THE GENUS ACTINOMETRA. 85

cular opening. So far as I know, these openings are tolerably constant on the dorsal
surface of the radial pentagon of Ant. rosacea ; but the five depressions corresponding to
them on the ventral surface of the centrodorsal piece are very variable in the distinct-
ness of their development ; and Dr. Carpenter has found that in some cases they may be
absent altogether \

This last condition, in which there are no radial depressions (q) on the ventral surface
of the centrodorsal piece, appears to be the normal one in Ant. celtica, in which I have
rarely found any traces of such depressions (PI. IV. fig. 2). The margin of the central
opening is usually almost circular, though sometimes bluntly stellate as in Ant. rosa
(fig. 15) ; at the same time, the five openings ( Q) upon the dorsal surface of the radial
pentagon are but little developed or even entirely absent. In PI. IV. figs. 3, 7, they are
represented as present in the small variety and absent in the large one ; but I have
sometimes found exactly the reverse to be the case.

"We shall find the same variability in the presence or absence of these openings in
Actinometra, not only in different individuals of the same species, but in the same indi-
vidual. This fact shows that too much reliance must not be placed on the presence,
absence, or difference in size of similar openings in the calyx of the fossil Crinoids
(the interradial " Li'icken " in Cupressocrinus, for example) as characters of any
systematic value. The absence or slight development of these openings in Ant. celtica is
principally due to the fact that the inner margin of the dorsal surface of the first radials
is not notched but straight, the radial axial furrow not being continued so far towards
the dorsal surface as in Ant. rosacea ; and also that processes grow inwards from the two
sides of the dorsal end of each of the five spout-like rays of the rosette, so that the
lumen of the canal it encloses becomes much diminished ; while in some cases similar
processes are put forward from the margin of the first radial, which unite with the
others so completely as eutirely to obliterate the lumen of the radial axial canal, and
thus form its dorsal boundary.

(§ 58) The dorsal aspect of the pentagonal base of the calyx of Actinometra is by no

1 Schliitcr (op. cit. p. 37) has proposed a division of the (fossil) Comatulce into two groups, characterized as fol-
lows : — (1) Centrodorsal with no radial pits and a round " Nahrungscanal " (central opening) ; and (2) Centrodorsal
with radial pits and a quinquepartite opening.

These characters, however, are far too uncertain to be of any systematic value. For example, Schliiter himself
notes the absence of the " Eadialgruben " in his own specimen of Solanocrinus scrobiculatus, while they were present
in one examined by Quenstedt (Echinodermen, p. 179). I have some specimens of Ant. celtica answering to the
first, and others to the second of the above definitions: and although most specimens at Ant. rosacea would be classed
in group (2), yet individuals with a pentagonal or even quinquepartite opening, ?»tZ no radial jiits, are not uncommon.
This last condition is very common among the ' Challenger ' Antedons. In fact, the radial pits of Ant. rosacea and
Ant. celtica arc peculiar to these species, and not always present even in them. In no other recent Comatula have
I found any thing exactly like them. They are not parts of the generally concave surface of each radial area, but
have distinct peripheral borders marking them off from these surfaces, and corresponding to the openings of the radial
axial canals enclosed within the spout-like processes of the rosette. In Act, pectinata, Act. polymorpTia, var. 2, and
in a new ' Ch.illeii' ' \edon there is a distinct pit at the central end of each radial area, which is merely the

deepened termination of a depression occupying its median line. Its nature (sects. 45 & (!1 ) is essentially the same as
that of the radial pits of Ant. rosacea, but its appearance is very different. Hence I cannot corroborate Sehluter's
statement that " Slanche Arten" possess '• Eadialgruben." He ouly describes them in ~> out of his 11 fossil species,
i besides the Solaru crinns scrooieidatus examined by Quenstedt.

12*

86 ME. P. H. CAEPENTEE OX THE GENUS ACTINOMETRA.

means of such a simple nature as it is in Ant. celttca (PL IV. figs. 3, 7), where it con-
sists solely of the five adjacent dorsal surfaces of the component radials. These are
somewhat elevated in the centre hut fall away towards the sides, where they are sepa-
rated from one another by slight furrows, corresponding in position with the five inter-
radial ridges on the ventral surface of the subjacent centrodorsal piece (PL IV. fig. 2, i.e).
In Actinometra, however, these dorsal interradial furrows are very marked ; but they are
not usually visible on the dorsal aspect of the radial pentagon, as they are occupied by
five long processes which radiate outwards from the angles of the central vacuity in
which the rosette lies (PI. V. fig. 3 ; PL VI. figs. 4, 9, 13, 21, S).

The presence of these rays of the basal star introduces an element of considerable
complexity into the dorsal aspect of the pentagonal base of Actinometra ; and its nature
will be best understood if we commence with the study of its component pieces in
the large Act. robusta. The dorsal face of each first radial of this species (PL V.
fig. 12) is slightly convex, so as to fit into the somewhat depressed radial area correspond-
ing to it on the ventral surface of the centrodorsal piece (PL V. fig. 14, r.ar). The
centre of its inner margin is, as in Ant. rosacea, marked by a deep notch, which indicates
the position of the axial radial furrow occupying the median line of the internal face
(fig. 10, a.r.f.). The latter is converted into a canal by the union of its inflected edges
with those of one of the radial spout-like processes of the rosette (figs. 12, 13, p), in the
manner already described by Dr. Carpenter for Ant. rosacea.

The central notch on the inner margin of the dorsal face thus becomes a round opening
(fig. 12, Q), similar to that seen in Ant. rosacea (PL IV. fig. 1G, Q). A bristle passed
through this opening towards the ventral side, therefore, will follow the course of one of
the axial radial canals, in which its lower end is concealed by the spout-like radial
process of the rosette (PL V. figs. 12, 13, I). On the ventral side of the rosette the
radial axial canal is incomplete, as the furrow on the internal face is only partially
bridged over by the calcareous processes which extend themselves from its sides to meet
the rosette ; the bristle which lies in the furrow is therefore visible here and there through
the openings in the network {cm) formed by the inosculation of these processes (figs.
10, 13, 1). This is best seen in fig. 13, which is a view of two radials from within, together
with that portion of the rosette which corresponds to and is united with them; and also
in fig. 10, which represents the internal face of a single radial, from which the portions
of the rosette that are normally united with it have been removed, so that the whole
of the internal face is exposed. The bristle I is seen to lie in the deep furrow between the
two raised edges of the apertures {x1, y) of the central canal, and to pass upwards under
the network extending from the ventral half of the internal face, where it follows the
course of the axial radial furrow and emerges on the ventral aspect of the radial. The
furrow in which it lies is here continued into the numerous irregular furrows of the
ventral face which converge towards the centre of its inner margin (PL V. fig. 11).

Just above the dorsal surface of the radial, the axial furrow occupying the median
line of its internal face gives off a large horizontal diverticulum into the substance of its
calcareous tissue, which extends outwards for some distance between the central canal
and the dorsal surface of the radial (fig. 10, r.c') ; and, like the axial furrow or canal as

ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA. 87

it is in the natural coudition when the rosette is in situ, encloses a dorsal extension of the
body-cavity or ccelom. I have seen no trace of these diverticula in any other species of
Comatula that I have yet examined ; but they are very large and well marked in each of
the five first radials in Act. robusta.

(§ 59) The furrows which occupy the median line of the ventral and internal faces of
the first radials thus terminate in Act. robusta (PI. V. fig. 12) precisely as in Ant. rosacea
(PL IV. fig. 16), by five large openings (Q) on the dorsal aspect of the radial pentagon,
which are closed in the natural condition by the ventral surface of the centrodorsal plate
on which the radial pentagon rests.

In Ant. rosacea the course of the slightly marked interradial farrows which pass down
from the ventral aspect of the radial pentagon into the peripheral portion of the central
calcareous network is terminated interiorly by the five triangular interradial processes of
the rosette ; for the apices of these processes unite with the two members of every pair of
contiguous radials, just between the two adjacent apertures of their central canals (PI. IV.
figs. 3, 7, 10, o).

In Actinometra, however, the interradial furrows both are more marked on the ventral
surface of the radial pentagon, and, like the radial ones, become converted into canals,
terminating by five openings upon its dorsal aspect.

In Ant. rosacea the edge which separates the lateral and dorsal faces of each first
radial is tolerably sharp and straight (PL IV. fig. 12 b, c); but in Act. robusta it is
somewhat truncated (PL V. figs. 10, 12, 13), so that when the lateral faces of two radials
are in apposition a deep interradial furrow appears along the line of union of their dorsal
surfaces (fig. 12, a.i.f). In the middle of the inner margin of the floor of this furrow is
a notch similar to that marking the centre of the inner margin of the dorsal face of each
single radial, both in this species and in Ant. rosacea (PL IV. fig. 12 b, Q'), except that
two radials take part in its formation instead of only one. This notch marks the con-
tinuation towards the dorsal surface of an interradial furrow from the ventral aspect of
the pentagonal base.

The edges between the internal and lateral faces of each first radial are truncated in
the same way as those between the dorsal and lateral faces (PL V. figs. 10, 13). In the
natural condition, therefore, when the lateral faces of all the radials are in apposition
with one another in pairs, there are five axial interradial furrows alternating with the
radial ones, which occupy the median lines of the internal faces. The ventral portions of
these, as of the axial radial furrows, are partially bridged over by the inosculating
calcareous processes which extend themselves towards the ventral aspect of the rosette
from the internal faces of the five first radials, so that a bristle passed along their course
is only partially visible (PL V. fig. 13, II).

These superior portions of the axial interradial furrows are in free communication,
both laterally, with the radial furrows occupying the intervals between them, and cen-
trally with the remaining spaces of the calcareous network, of which system these two
sets of furrows form the peripheral part. Interiorly, i. e. towards the dorsal surface, each
of these axial interradial furrows passes between the two outer lips of the adjacent aper-
tures (x, x') of the central canals of two contiguous radials along the line of union of

8S MS. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

which the interradial furrow is situated. The outer lips of these apertures, like the inner
ones (PI. V. fig-. 10), are raised and applied to the similarly inflected edges of the five
spout-like interradial processes of the rosette, so that the furrow lying hetween the aper-
tures becomes converted into a complete canal. A bristle, therefore, which lies in the course
of this furrow (PI. V. figs. 12, 13, II) is concealed by the interradial process (o) of the
rosette. The dorsal end of the latter unites with the margins of the notch described above
at the central end of the dorsal interradial furrow, so as to produce a roundish interradial
opening on the dorsal aspect of the pentagonal base, through which the bristle passed
along the axial interradial furrow emerges from its concealment beneath the interradial
process of the rosette (figs. 12, 13, II). The manner in which these openings are closed
in the natural condition by the central ends of the rays of the basal star will be best
described further on.

(§ 60) In the condition and relative inclination of their dorsal and internal faces, the
first radials of Act. robusta are more like those of Ant. rosacea than those of any other
of tin- various Actinometree which I have examined. In Ant. rosacea the ventral sm'face
of the eentrodorsal piece (PL IV. fig. 15) is almost flat, as the five radial areas into which
it is divided lie nearly in a horizontal plane; and I he corresponding dorsal surfaces of
the five first radials are likewise horizontal, and form an angle of but little more than 90°
with the internal faces (PI. IV. fig. 12 b, c). In Act. robusta this angle becomes more
obtuse, so that the dorsal surfaces of the radials are somewhat inclined to the horizontal
plane (PL V. figs. 10, 13) ; and, in correspondence with this, the radial areas on the
ventral surface of the eentrodorsal (fig. 11, r.ar.) have a slight downward and outward
slope between their central and peripheral margins, so that the whole surface rises very
gradually from the circumference towards the centre.

This is also the case in Act. Solaris, in which the dorsal surface of the radial pentagon
slopes slightly downwards from its margin towards the opening of the central vacuity
in which the rosette is situated (PL V. fig. 3), so as to correspond with the gradual eleva-
tion between the circumference and centre of the ventral surface of the eentrodorsal on
which it rests (fig. 2).

Act. Solaris also agrees with Act. robusta in the fact that the sides of the dorsal inter-
radial furrow (fig. 3, a.i.f) which is produced by the truncation of the adjacent supero-
lateral edges of two contiguous radials are simple and straight, and not raised into
leaf-like folds, as in Act. pectinata (fig. 9 b) and Act. polymorpha (PL VI. figs. 9, 13, 21 ;
PL VII. figs. 1 d, 1 d, b.g).

In Act. Solaris there are none of the apertures which occur in Ant. rosacea and Act.
robusta, by which the axial radial canals open upon the dorsal surface of the radial
pentagon (PL IV. fig. 10, and PL V. fig. 12, Q). We have already seen (sect. 57) that they
may be absent in Ant. celtica (PL IV. fig. 3) ; and their absence in Act. Solaris is due to
the same cause as in this case, viz. to the want of a central notch on the inner margin
of the dorsal face of each first radial, and to the obliteration of the lumen of each canal
by the ingrowth of calcareous tissue from its sides.

In the closely allied Act. pectinata, however, these openings may be present (and not
improbably also in other specimens of Act. Solaris than the one which I have examined);

ME. P. II. CAEPENTEH ON THE GENUS ACTINOMETEA. 89

for the inner margin of the dorsal face of each first radial exhibits a slight median notch
(PL V. fig. 9 b, Q), though it is by no means so distinct as in Ant. rosacea (PI. IV.
figs. 12 b, 16) and Act. robusta (PL V. fig. 12).

In this species too the ventral surface of the centrodorsal plate (PL V. fig. 7) rises very
perceptibly from its circumference towards its centre ; and the dorsal face of each first
radial is very considerably inclined to the vertical internal face, the angle between the
two almost reaching 135° (PL V. fig. 9 b, c). Consequently, when the radial is viewed
from its dorsal side, the large projecting lips of the two apertures of its central canal are
seen below the central or inner edge of the inclined dorsal face (fig. 9 b, x, y). These
are not seen in a similar view of a first radial of Ant. rosacea, in which the inclination
of the dorsal to the internal face is very little over 90° (PL IV. fig. 12 b, c).

(§ 61) In Ant. rosacea and Act. robusta the slight convexities of the dorsal surfaces
of the first radials fit into the correspondingly slight concavities in the centre of the
radial areas on the ventral surface of the centrodorsal piece (PL IV. figs. 2, 3, 15, 16;
PL V. figs. 12, 11). In Act. pectmata, however, these areas are occupied by median
depressions, increasing somewhat in depth from their peripheral to their central ends
(PL V. fig. 7, r.ar) ■ but the dorsal faces of the first radials do not exhibit corresponding
ridges, for they have similar median depressions, which are also deepest at their central
ends (PL V. fig. 9 b, c, d.r.f).

When, therefore, the dorsal surface of the radial pentagon and the ventral surface of
the centrodorsal piece are in their normal state of apposition, they are separated from
one another along the median lines of the five radials by five cavities or radial spaces ;
these are largest at their central ends, and extend in a peripheral direction to open
externally by five small openings situated on the margin of the small centrodorsal
piece, beneath the radial pentagon which rests upon and extends considerably beyond it.
These "radial spaces " are seen in section in PL VIII. figs. 5-8, which represent parts
of foiu- sections selected out of a series that was cut through a decalcified calyx of
Act. pectmata.

The section represented in fig. 5 was cut across the angle of two radials {A, B) near
the edge of the centrodorsal piece, and the open outer ends of the radial spaces are cut
somewhat obliquely (r.s). Pig. 6 represents a section much nearer the centre, and the
closed inner ends of the radial spaces are seen just beneath the lower ends of the axial
radial canals (a.r.c), but not communicating with them. In fig. 7 two other spaces are
seen, cut almost longitudinally, as the section is one from the other side of the centre,
through the radii, C & E, almost in the direction of their axial nervous cords (»), beneath
which are the radial spaces (r.s) between the dorsal surfaces of the first radials and the
ventral surface of tbe centrodorsal piece. Lastly, in fig. 8, which represents a section
still further from the centre, and cut transversely to the direction of radius 1), the closed
central end of the corresponding radial space is seen, as in fig. 7, on the dorsal side of the
axial nervous cord (n); at either side of it (s) is the expanded lower end of one of the
axial interradial canals seen in fig. 7 (a.i.c).

The external medium which occupies these radial spaces between the radial pentagon
and the centrodorsal piece is only shut off from the dorsal portion of the coeloui en-

90 MR. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

closed within the radial pentagon, and from the centrodorsal coeloni, by the small
bony bars at their expanded central ends, which form the thickened inner or central
edges of the dorsal faces of the first radials (PL V. fig. 9 b, c.) They are slightly
developed in one of the varieties of Act. polymorpha, in which the radial areas of the
small centrodorsal piece (PI. VI. fig. 17, r.ar) and the corresponding dorsal faces of the
first radials (PI. VII. fig. I a, d) both exhibit median depressions, which gradually
increase in depth from their peripheral to their central ends. In both these cases the
centrodorsal piece is relatively very small and by no means conceals the first radials,
while its ventral surface rises very considerably from the circumference towards the
centre (PI. V. fig. 7, and PI. VI. fig. 17). The meaning of these radial spaces is to me
quite obscure. [n no other Comatula have I found any thing at all comparable to them
except in Act. robusta, where the axial radial canals give off horizontal diverticula (PL V.
fig. 10, r.c'), which extend outwards in a peripheral direction in the substance of each
first radial just beneath its dorsal surface. These, enclosing diverticula of the dorsal
portion of the ccelom, are, of course, in indirect communication with the external
medium, while the radial spaces in Act. pectinata are altogether outside the substance of
the first radials, communicate directly with the exterior, and are completely shut oil' from
the dorsal ccelom. There is, therefore, scarcely any resemblance between the two sets of
cavities, although they occupy very nearly similar positions, i. e. between the centrodorsal
piece and the whole or the greater part of the mass of the first radials.

In some species of the fossil Apiocrmus, however, cavities similar to the radial spaces
in Act. pectinata appear to exist between every pair of contiguous basals and the first
radial, which rests upon them and alternates with them in position. As the basal circlet
is generally regarded as comparable to a stem-segment, it is evident that the posi-
tions of these cavities in Act. pectinata and in Apiocrmus respectively are homologous
with one another. In Apiocrinus rotundas five lateral openings were discovered by
Miller ' on the circumference of the body, " in or between the lateral surfaces of the
joints of the pelvis (basals) and the insertion of the first costal (radial) joints," which in
one case he thought he was able to trace as a canal or perforation " passing through the
joint of the pelvis into the space between it and the costal joints, extending perhaps
thence into the perivisceral cavity" (i.e. into the dorsal division of the body-cavity).

Miller supposed these to be the openings of oviducts leading to an ovary situated in
this dorsal ccelom, just in the same manner as the five openings on the ventral surface
of the centrodorsal piece of Glenotremitcs were (till lately) regarded as genital openings,
although the genital glands of all the recent Crinoids with which we are acquainted are
situated in the arms and pinnules.

Similar openings to those seen by Miller in Ap. rotundas have been described in Ap.
obconicus by Goldfuss 2, who also supposed them to lead into the body-cavity. This is,
however, certainly not the case with the homologous openings in Act. pectinata.

The interarticular pores in the upper part of the stem of Peiitacrinus are also homolo-
gous with the external openings of the radial spaces in Act. pectinata. They are the

1 Op. tit. p. 31. 2 Petref. German, p. Is7, Taf. vii. fig. 5, a, b, c.

ME. P. H. CAEPEXTEE ON THE GENUS ACTINOMETEA. 91

openings of spaces between the successive segments, which are similarly situated, with
regard to the radial symmetry of the animal, to the radial spaces in Act.pectluala, viz.,
in the direction of the radii ; and they are produced in the same way, by the apposition
of two grooves radiating outwards from the centre of each stem-segment, which are
largest at their central ends and shallowest towards the periphery.

(§ 62) In Ant. rosacea and celtica (PI. IV. figs. 3, 7, 16), and in Act. rohusta and Solaris
(PI. V. figs. 3, 10, 12, 13), the sides of the interradial furrows (d.i.f) on the dorsal surface
of the radial pentagon are simple and straight ; but in Act. pectinata that portion of the
dorsal surface of each first radial which is next to its truncated lateral edge is raised
into a sort of curved ridge or fold (PI. V. fig. 12 b, b.f), so that in the natural con-
dition of mutual apposition of the five first radials the dorsal interradial furrows become
somewhat lancet-shaped. They correspond in position with the basal grooves on the
ventral surface of the subjacent centrodorsal piece (PI. V. fig. 7, b.g), and in the cavity
formed by the apposition of the edges of these two grooves lie, as will be subsequently
seen, the five rays of the basal star.

The first radials of Act. polymorpha are very similar to those of Act. pectinata. Those
of variety 1 (PI. VI. fig. 12) are like those of the type (PI. VII. tig. 1), except in the
simpler condition of their ventral surface, which is far less marked by secondary ridges
and furrows than is the case in the type (PI. VI. fig. 5). In the other three varieties,
the first radials of which resemble one another very closely, this sculpturing of the
ventral surface is even more marked than in the type (PI. VI. fig. 23 ; PI. VII. fig. 1 c).
The angle between the dorsal and internal faces is considerably less in the type
(PL VII. fig. 1 a,d) and in var. 1 than in varieties 2, 3, and 1, the first of which
resembles Act. pectinata in the presence of a median depression of the dorsal face
(PL VII. fig. 1 a,d, d.r.f), which corresponds with a similar depression along the
median line of the radial areas of the small centrodorsal piece (PL VI. fig. 17, r.ar).
This dorsal interradial furrow does not exist in varieties 3 and 1, nor in var. 1 (fig. 13),
while there is a trace of it in some specimens of the type, but not in others (figs. 1, 9).
In like manner the development of the openings of the radial axial canals on the
dorsal surface of the pentagonal base, which are so large in Ant. rosacea and in Act.
rohusta (PL IV. fig. 16 ; PL V. fig. 12, Q), is in Act. polymorpha extremely variable. In
two specimens of the type (PL VI. figs. 1, 11) they are entirely absent, as in Act.
Solaris (PL V. fig. 3) ; in another the inner margin of the dorsal face of each first radial
exhibits a slight median notch (PL VI. fig. 9 ; PL VII. fig. la, Q'), which would be
completed into an opening by the apposition to it of the end of one of the radial
spout-like processes of the rosette.

In variety 1 this notch is fairly marked, and five small openings are consequently
visible around the central vacuity, on the dorsal surface of the pentagonal base (PL VI.
fig. 10, Q). In varieties 2 (PL VII. fig. la) and 3 it is somewhat more distinct; and
in var. 1 it exists iu three of the first radials, but not in the other two, so that there
arc only three openings on the dorsal surface of the pentagonal base (PL VI. fig. 21, Q).

The extent to which the basal folds are developed at the sides of the dorsal inter-
radial furrows is also very variable in Act. polymorpha. We have seen that, although

SECOND SERIES. — ZOOLOGV, VOL. II. 13

92 ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

absent in Act. Solaris (PL V. fig. 3), they are well marked in the closely allied Act. pec-
tinata (PI. V. fig. 9 h, h.f). In two specimens of the type of Act. polymorpha (PI. VI.
figs. 4, 11) they are absent altogether, while in a third they are very well marked (fig. 9,
b.f), as also in each of the varieties, three of which are represented in PI. VI. figs.
13, 21, and PI. VII. fig. 4 a. In all these cases the borders of the interradial furrows
on the dorsal surface of the pentagonal base, which are produced by the apposition of
the truncated superolateral edges of every pair of contiguous radials, assume a leaf-like
appearance, owing to the presence of the folds at their sides. The precise shape of these
leaves, which is different in the type and in all the four varieties, corresponds very
closely with the shape of the basal grooves on the ventral surface of the centrodorsal
piece, with which they also correspond in position. They further resemble them in
the fact that they are entirely devoid of the pigment which is so abundant on the other
parts of the surfaces of the radial pentagon and centrodorsal piece ; so that when these
last are separated from one another, the dorsal interradial furrows on the pentagonal
base, like the basal furrows on the centrodorsal piece, stand out sharp and distinct as
five white leaflets on a dark-brown background. They are best marked invar. 4 (PL VI.
fig. 24, b.f), in which the basal folds of every pair of contiguous radials are rather widely
separated from each other about the middle of their length. This is also the case,
though to a less extent, in var. 1 (tig. 13), where, as in var. 4, the dorsal interradial
furrows correspond very closely in shape with the basal grooves on the centrodorsal
piece (fig. 15).

This is particularly distinct in the specimen of the type represented in PI. VI.
figs. 8, 9, in which one of the basal grooves is much shorter than the rest, and does
not reach the margin of the centrodorsal piece. The basal folds at the sides of the
dorsal interradial furrow corresponding to this aborted groove are also imperfectly
developed, so that the borders of its outer end are simple and straight ; as they are
throughout the whole course of the furrows in Act. solans (PL V. fig. 3) and Act. robusta
(PL V. fig. 12).

This last condition may also occur in the type of Act. polymorpha (PL VI. figs. 4, 11) ;
and in correspondence with it the basal grooves on the centrodorsal piece are simple and
almost parallel-sided (PL VI. figs. 3, 10, b.g), just as in Act. Solaris and Act. robusta
(PL V. figs. 2, 14).

This correspondence in the appearance of the dorsal interradial furrows and basal
grooves which is also seen in Act. polymorpha, var. 2 (PL VI. fig. 17; PL VII. fig. 4 a),
is not, however, an invariable one; for in Act. pectinata the basal folds are very well
marked (PL V. fig. 9 b, b.f), and the dorsal interradial furrows, therefore, leaf-like in
appearance, as in most specimens of Act. polymorpha (PL VI. figs. 9, 13, 24). The basal
grooves, however, on the ventral surface of the centrodorsal piece are narrow and
parallel-sided (PL V. fig. 7, b.g), just as in the allied species Act. Solaris and Act. robusta
(PL V. figs. 2, 14).

The external or distal faces of the first radials of Act. polymorpha differ not a little
from the corresponding faces in Ant. rosacea and Ant. celtica (PL IV. figs. 4, 6, 8 , 12 a,
11, 17) ; in which, especially in the latter, the fossse (f) for the attachment of the muscles

ME. P. H. CAEPEXTEE OX THE GENUS ACTIXOMETEA. 93

are very large, and considerably more extensive than those which lodge the interarticular
ligaments (I/).

In Act. polymorpha (PI. VI. fig. 1; PI. VII. figs, lb, 4b), however, the muscular
fossae are very small, being best developed in var. 1 (PL VI. fig. 12,/); while the fossae
(h) lodging the interarticular ligaments are very extensive, and separated by the down-
ward continuation of the intermuscular furrow (ft), which reaches the dorsal margin of
the opening of the central canal (c.c). The external faces of the first radials of varieties
2 (PI. VII. fig. 4 b), 3, and 4 resemble one another, but differ from the corresponding
faces in the type (PI. VII. fig. lb) in being somewhat higher in proportion to their
width, and in the fact that the fossae (j) lodging the elastic ligaments are relatively
smaller, not extending so far into the lower or dorsal angles of the face as is the case in
the type.

(v.) The Basals.

(§ G3) We have already seen that all the older observers regarded Comatula as devoid
of those five pieces resting upon the top segment of the stem to which, in the other
Crinoids, MuTler gave the name of " basals ; " and it was not until Dr. Carpenter l dis-
covered the extraordinary metamorphosis undergone by the embryonic basals of Coma-
tula (Antedon) rosacea and then transformation into the "rosette," that the existence
of basals, although internal and concealed in the adult animal, was recognized.

The rosette of Ant. rosacea and Ant. celtlca (PL IV. figs. 3, 7, 16, H) is a peculiarly
shaped circular plate, occupying the dorsal half of the central cavity in the pentagonal
base of the calyx, which lies much nearer to the dorsal surface of the pentagonal base
in the latter species than in the former.

A normal rosette consists of a disk perforated in the centre with ten rays proceeding
from it. Pive of these rays (PL IV. tig. 13, o') are short, triangular in form, and nearly
flat, and their position is interradial, as they are directed to the sutures between the five
radials, their apices joining the contiguous pairs of these just between the two adjacent
apertures (x, x') of their central canals.

Alternating with these five interradial processes of the rosette are five radial ones
(fig. 13, p), each of which has parallel margins inflected on its ventral aspect in such
a manner as to form a groove ; while the process itself is so curved towards its dorsal
aspect that this groove reaches the periphery of the rosette, and then terminates
abruptly as if truncated.

The inflected margins of each of these five radial or, as Dr. Carpenter has called them,
" spout-like " processes of the rosette are applied to the similarly inflected margins of the
dorsal half of the axial radial furrow, lying between the two apertures of the central
canal on the internal face of each first radial (fig. 12 c, x', //). In this manner a com-
plete radial axial canal is formed, which, as we have already seen, terminates on the
dorsal surface of the radial pentagon by the opening Q (fig. 10), or becomes closed
before it reaches the dorsal surface by the union of ingrowths developed from its walls.

Besides this very intimate union between the peripheral portion of the rosette and the

1 Phil Trans. 1865, pp. 744, 745.

13*

9i MR. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

internal faces of the first radials, its central portion is also frequently connected with
the radial pentagon by delicate processes, which sometimes sprout forth irregularly
from the inner margins of the component pieces of the latter ; but sometimes form a
more regular ingrowth, which considerably contracts the central space on the ventral
aspect of the disk, and becomes continuous with an annular projection from the ventral
face of the rosette.

(§ 64) Before attempting to understand the complicated condition of the basals in
Actinometra, it Avill be well to study the mode in which the embryonic basals of Ant.
rosacea become metamorphosed into the rosette, as described by Dr. Carpenter, from
Avhose memoir the following account is principally taken.

In the young animal each basal is a flattened irregularly pentagonal plate, the apex of
which lies between a pair of radials that partially rest upon it. On the ventral surfaces
of the basal plates lie the five primary basal cords 7\ IV, X, Y, Z, proceeding from the
angles of the quinquelocular organ. Each of these divides into two branches, Vx, V2,
. * . . Zu Z2, the secondary basal cords, which pass on to the ventral faces of each pair of
contiguous radial plates, c. <j. X, and 1\ to one radial, T2 and Xx to the next, and so on.

Both basals and radials gradually become much thickened by an endogenous extension
of the calcareous network, which takes place in such a manner that the basal cords
come to lie in furrows channelled out on the ventral surfaces of the plates. By a further
endogenous growth of the radial plates these furrows are converted into canals (the
" central canals " of Johannes Mtiller), which at first lie close under the ventral surfaces
of the plates, but come gradually, by a continuation of the same process, to lie in then-
central axis.

In the basals, however, this process of endogenous growth is followed by one of
absorption; for the cribriform film of which each basal is originally composed, and
which still forms its external layer, now undergoes absorption, especially in its central
portion, where it covers in the dorsal side of the primary basal cord ; so that the central
space left by the incomplete union of the proximal ends of the five embryonic basal
plates is extended on its dorsal aspect into five broad rays, though on its ventral aspect,
where it is bounded by the last-formed portion of the endogenous reticulation, it shows
no corresponding increase. It is this last-formed ventral portion which persists in the
adult as the five triangular intcrradial processes of the rosette (PI. IV. figs. 3, 7, 13,
16, d).

The formation of the five radial or spout-like processes is somewhat more complicated.
The removal of the external layer in the centre of the dorsal aspect is carried so far as
to leave nothing but a kind of thickened margin along those sides of the plate which are
received between the first radials ; and by an extension of the same process along the
median dorsal line of each plate as far as its salient angle, so as completely to remove
the terminal portion of its inferior or dorsal layer, its two lateral portions become sepa-
rated from each other at their distal ends, and remain as small curved processes extend-
ing outwards. Those of every two contiguous basals now unite to form a sort of ray
curving towards the dorsal aspect; and this is the rudiment of one of the five radial or
spout-like processes of the rosette, the shape of which becomes much more strongly pro-

ME. P. H. CAEPENTEB ON THE GENUS ACTINOMETEA. 93

nounced with the subsequent increase of its size (PI. IV. fig. 13, p). The rosette is thus
essentially formed at the expense of the secondary or ventral layer of the original basals,
the ends of the curved spout-like processes being- the sole residue of their primary or
dorsal layer ; and since, by the removal of the median portion of that layer in each plate,
the primary basal cords are left bare upon their dorsal aspect, they now pass from the
angles of the quinquelocular organ into the central canals of the first radials, on the
inferior or dorsal side of the calcareous skeleton which occupies the base of the calyx ;
instead of reaching them by passing, as they did in the first instance, along its superior
or ventral face or, as at a later period, through the middle of its substance.

Each of these primary basal cords, X, Y, &c, which are thus interradial in position,
divides into two branches, Xu X,, Yx, Y2, &c, towards the periphery of the rosette, on
the dorsal surface of which it rests. These branches lie in the shallow channels which
mark the union of the base of each interradial triangular process (fig. 13 1), </) with the
two curved lateral processes above mentioned, each of which unites with a corresponding
process from the adjacent basal to form one of the five spout-like rays Qj) of the rosette.
The apex or peripheral end of each triangular process is directed to the suture between
two contiguous radials (figs. 3, 7, 16, d), to which it is attached just between the two
adjacent apertures (a; .?') of their central canals. Into these canals pass the secondary
basal cords Xu X2, one into each of the two contiguous radials, so that one lies on each
side of the interradial process of the rosette.

(§ 65) As a general rule, this process, both in Ant. rosacea and in Ant. celtica, is short,
triangular, and slightly curved towards the ventral side. It is not always so, however,
for I have frequently met with specimens of Ant. rosacea in which one or more of the
interradial processes of the rosette, after bending for a short distance towards the ventral
side, turns suddenly downwards, and extends towards the dorsal surface of the radial
pentagon. At the same time the parallel margins of each of these abnormally developed
processes are so inflected towards the dorsal surface as to form a narrow interradial
spout-like process. This is so applied to the projecting and similarly inflected outer
edges of the adjacent openings of the central canals {x, x1) in two contiguous radials
as to convert the interradial furrow lying between them into a complete axial inter-
radial canal, precisely similar in character to the radial axial canals already described
(§§52,57).

In one case which I have met with, four out of the five interradial processes of the
rosette were of this character. In the rosette represented in PL IV. fig. 13, only two of
the interradial processes (o) are long and spout-like, the other three (o') are short and
triangular, like those of a normal rosette.

This abnormal condition of the interradial process of the rosette of Ant. rosacea is of
considerable interest, as it is the normal one in Actlnometra and in many species of
Antedon.

Not only the interradial, but also the radial processes of the rosette of Ant. rosacea
may exhibit departures from their usual shape ; for the removal of the primary or
dorsal layer at the salient angle of one or more of the five embryonic basals may be
incomplete, so that the ends of the curved rays of the rosette exhibit lateral processes,

96 MR. P. H. CAEPENTER ON THE GENUS ACTINOMETRA.

which are the remains of the upper margins of the primitive hasal plates on which
the first radials rested. Occasionally the apex of the original hasal is left unabsorbed,
so that the two lateral curved processes which remain after the removal of the primary
external layer along the median line of each "plate remain in connexion with one another;
as is seen in the bottom part of the rosette represented in PL IV. fig. 13. The trian-
gular iuterradial process (o'), which is developed from a secondary calcareous deposit
on the ventral side of the original basal, has here become more or less completely united
with these primary bars connecting the two lateral portions of the basal. The latter
retain their primitive relation to the first radials, for they remain united with them
along the inner margins of their dorsal faces (fig. 10, b.b) ; and as they partially cover
in the secondary basal cords, Xu X2, &c., on their dorsal aspect before they enter the
central canals of the first radials, I will call them the " basal bridge " (PL IV. fig. 13,
b.b).

This basal bridge is well seen in situ in the specimen of Ant. rosacea represented in
PL IV. fig. 10. It is remarkably well developed, being nearly as distinct as in Actino-
melra (PL V. figs. 3, 12, and PL VI. figs. 1, 13, 21, b.b), in which its presence is normal,
and not al (normal, as in Ant. rosacea. It is also slightly developed in the specimen of
Ant. celtica represented in PL IV. fig. 3 ; but in fig. 7 no trace of it is visible.

(§ G6) This tendency to an incomplete metamorphosis of the embryonic basals of Ant.
rosacea, and consequently to the abnormal persistence of a more embryonic condition
than usual, is of considerable interest, because in Actmometra and in many Anledons
a basal bridge, representing the apex and unabsorbed margins of the embryonic basal
plates, is normally present (PL V. figs. 3, 8, 12, and PL VI. figs. 4, G, 13, 18, 19, 22, 24,
b.b). While at the same time, as already mentioned, the iuterradial processes of the
rosette (o), which are developed from a secondary or ventral layer, are large and spout-like,
as is abnormally the case in Ant. rosacea, and acquire a connexion with the remains of the
primary or dorsal layer which forms the basal bridge. The complicated rosette thus
constituted also becomes united with the large, more or less spindle-shaped, rays of the
basal star (S), the origin of which, as will be subsequently seen, is totally different from
that of the rosette.

A single " compound basal," as it may be called, of Actinometra, thus consists of two
distinct elements — (i) the incompletely metamorphosed embryonic basal, and (ii) a single
ray of the basal star. Its position is iuterradial, as it occupies the space enclosed between
the apposed edges of the basal furrows on the ventral surface of the centrodorsal piece,
and of the iuterradial furrows ou the corresponding dorsal surface of the radial pentagon
(PL VI. figs. 4, 13, 24,#).

An isolated compound basal which is thus constituted, when seen from its dorsal side
(PL V. fig. 8 b, PL VI. fig. 22 b), shows : — (i) more or less of the calcareous network {cm)
which unites the ventral surface of the rosette to the internal faces of the first radials ;
(ii) a large iuterradial spout-shaped process (o) ; (iii) two small, radial, curved processes
(//), extending outwards from the base of the iuterradial process, and representing the
unabsorbed lateral portions of the primary layer forming the embryonic basal plate.

ME. P. H. CARPENTER ON THE GEXUS ACTINOMETRA. 97

(iv) The basal bridge (b.b), consisting of two calcareous bars, that represent the
unabsorbecl peripheral margins of the embryonic basal on which two first radials rested.
They extend towards one another from the outer ends of the small radial processes, until
they meet at a point that represents the apex of the embryonic basal, and is situated
on the dorsal side of the peripheral end of the interradial process (o), developed from the
secondary or ventral layer, which becomes united with the basal bridge.

(v) The ray (S) of the basal star, which is joined to the interradial process and to the
basal bridge, along the line of union of the two primary bars constituting the latter with
one another and with the secondary interradial process, i. e. at the apex of the embryonic
basal. The development of this ray is quite different from that of either the primary or
the secondary portions of the compound basal. It is really a tertiary structure, being
nothing more than a deposition of calcareous material in the substance of the connective
tissue of the synostosis between the centrodorsal piece and the radial pentagon.

(vi) At the sides of the interradial process (o), bounded laterally by the radial pro-
cesses (p), and externally by the bars of the basal bridge {b.b), are two large apertures,
*u #3j >/\, 1/-j, &c., in each compound basal. Through these apertures pass the secondary
basal cords, Xx, X,, Yl} Y,, &c. (PI. VIII. fig. 3), which result from the bifurcation of the
primary cords, X, Y, Z, proceeding from the angles of the quinquelocular organ. The
two secondary cords lie in the depressions on the dorsal surface of the compound basal,
between the central ends of its radial and interradial processes. They then pass outwards
through the apertures (a\, a\, &c.) beneath the bars of the basal bridge, and enter the
adjacent openings (%, a/, &c.) on the internal faces of the two contiguous first radials1,
which contribute to form the dorsal interradial furrow occupied by the single fusiform
ray (S) of the corresponding basal.

The ventral surface of each of these rays of a compound basal (PI. V. fig. 8a; PI. VI.
figs. 6, 18, 22 a) is not flat, like the dorsal surface, but occupied by a prominent median
ridge, so that the ray is triangular in section. This ridge does not extend quite to the
central end of the ray, which is occupied by a considerable depression (s), forming the
peripheral end of the groove contained in the spout-like interradial process (o). In the
natural condition, when the basals are in situ and in connexion with the radial pen-
tagon, the inflected edges of this process unite with those of the axial interradial furrow
to form an axial interradial canal. This terminates on the dorsal surface of the radial
pentagon by a small opening situated at the central end of the dorsal interradial furrow
(PI. V. fig. 12, II), in which furrow the tertiary element (&') of the corresponding com-
pound basal is received. The depression (s) at the central end of the ray (PI. V. fig. 8 a ;
PI. VI. figs. G, 18, 22 a) lies over this opening, and thus forms a blind end to the axial
interradial canal (PI. VIII. figs. 3, 5, 7, a.i.c ; fig. 8, s)— precisely in the same manner
as the depressions (q) on the ventral surface of the centrodorsal piece of Ant. rosacea
(PI. IV. fig. 15) receive the blind ends of the axial radial canals which open on the
dorsal surface of the radial pentagon by the five large openings, Q (PI. IV. fig. 16).

A view of a single compound basal does not, of course, show one of the large and

1 In PL V. fig. 13, four of these openings are seen on the internal faces of the two contiguous first radials,
viz. x', y, y\ z.

98 MR. P. II. CARPENTER ON THE GENUS ACTINOMETRA.

spout-like radial processes of the rosette ; for each of these is a composite structure,
formed by the apposition of two of the small curved lateral processes of contiguous basals
(PL VI. rig-. 22, p'). This is seen in PI. VI. rigs. 6, 18, 19, particularly in the last two ; for
the union of the adjacent lateral processes (//) of the two contiguous basals which are
there represented, is seen to be incomplete, so that a slight fissure is visible along the
median line of the dorsal surface of the composite radial process (fig. 19, p). The peri-
pheral end of this radial process is united to those of the interradial processes (o) at its
sides by the bars of the basal bridge (b.b). Their central ends arc also united around the
opening of the rosette (PI. VI. figs. 4, 19, 24, II) ; but their median portions are separated
by the two apertures (a?2, yx) by which the two adjacent secondary basal cords ( X>, li) pass
out under the bars of the basal bridge, to enter the two openings of the central canal on
either side of the axial radial furrow on the internal face of the first radial (PL V.
figs. 9 c, 10 ; PL VII. figs. 1 a, 4 a, a/, y), with the inflected inner edges of which the
radial spout-like process (p) unites.

The openings (%u y._) by which the other branches (Xu Y2) of the two primary basal
cords pass outwards to reach the central canals of the other two radials corresponding to
these two basals are best seen in a dorsal view as shown in PL VI. fig. 19. This also
shows the two outer lateral processes (//) of these united basals, which would naturally
unite with those of the two next contiguous basals, one on each side, to form two more
radial spout-like processes.

(§ G7) The tertiary elements which form the rays of the basal star vary very con-
siderably in their shape and in the completeness with which they are developed, just as
do the interradial furrows on the dorsal surface of the radial pentagon in which they lie.
In Act. Solaris (PL V. fig. 3, d.i.f) these have no curved folds at their sides, and the rays
of the basal star (*S') are only imperfectly calcified rods, long and narrow, like the basal
grooves on the ventral surface of the centrodorsal piece into which they are received
(PL V. fig. 2, b.g). In Act. pectinata, however, although the basal grooves are long and
narrow as in Act. Solaris (PL V. fig. 7), yet the dorsal interradial furrows are widened by
the presence of large curved basal folds at their sides (fig. 9 b, b.f) ; and in correspond-
ence with these the tertiary basal elements (fig. 8, S) are much wider, and also far
more perfectly calcified (being solid throughout), than is the case in Act. Solaris.

In Act. robusta the central ends of these basal rays are wide and stout, and com-
pletely calcified as in Act. pectinata ; but their peripheral ends are much thinner, and
consist of a simple curved plate, which forms a sort of bridge over the dorsal interradial
furrow (PL V. fig. 12, S), the borders of which are straight, as in Act. Solaris (fig. 3, d.i.f),
and not marked by any lateral folds. The basal grooves on the ventral surface of the
centrodorsal piece are also simple and nearly parallel-sided (fig. 14, b.g).

In Act. polymorpha and its varieties the condition of the basal star varies extremely,
like that of the basal folds and basal grooves, the development and shape of which
exhibit a very close correspondence with the appearance of the rays of the basal star.

Thus, in that specimen of the type of Act. polymorpha in which the basal grooves are
narrow and parallel-sided, and all terminate well within the margin of the centrodorsal
piece (PL VI. fig. 3, b.[/), Avhile no basal folds are present at the sides of the dorsal inter-

ME. P. H. CARPENTER ON THE GENUS ACTIXOMETRA. 99

radial furrows (fig. 4, d.i.f), the rays of the hasal star are short and flattened, and do not
hy any means reach the angles of the radial pentagon (fig. 4, S). Their dorsal surface is
somewhat depressed along the median line, while the depression (fig. 6, s) at the central
end of the ventral surface which receives the Mind end of the axial interradial canal is
continued outwards in a peripheral direction somewhat further than in Act. pectlnata
(PI. V. fig. 8 a, s); and the median ridge which runs from its end to the apes of the ray
is less marked than in this species.

In another specimen of the type, however, in which hoth hasal grooves (PL VI.
fig. 8, b.g) and hasal folds (fig. 9, bf) are wide and well marked, the basal rays are
stout and thick, with a fairly distinct median ridge on their ventral surface. In fig. 8,
three of them are seen occupying their normal position in the hasal furrows on the
ventral surface of the centrodorsal piece, with which they are closely connected, while
the other two rays have remained in connexion with the rosette and radial pentagon.

In a third specimen of the type the rayS of the hasal star are very imperfectly
developed ; two only extend for any distance towards the angles of the radial pentagon
(PL VI. fig. 11, <$'), while of the other three little or nothing is to he seen. In this
specimen, as in the one first described, there are no basal folds, and the basal grooves
are parallel-sided and only imperfectly developed (fig. 10, b.g). It is also remarkable
for the fact that the absorption of the apex and outer margins of each embryonic basal
plate seems to have been very incomplete ; for the bars of the basal bridge are so wide,
and extend so far towards the centre from the inner margins of the dorsal surfaces of the
first radials, with which they are closely united, that they entirely conceal the apertures
in the compound basals (PL V. figs. 3, 8; PL VI. figs. 1, 6, 13, 18, 19, 22, 21, xu x,, yx,
y2, &c.) through which the secondary basal cords pass in order to reach the central
canals of the first radials. Consequently nothing is seen of the rosette in a dorsal view
of the pentagonal base but its central opening surrounded by a raised rim (PL VI.
fig. 11, r.o). In all the other figures, however (PL V. fig. 3 ; PL VI. figs. 1, 13, 21), these
apertures are large and distinct, every one being situated between a radial (p) and an
interradial process (o) of the rosette.

In Act. polymorpha, var. 1, both basal grooves (PL VI. fig. 15, b.g) and basal folds
(fig. 13, b.f) are well marked and somewhat lancet-shaped in form; the rays of the
basal star wdrich occupy the former are much flattened dorsally (fig. 22 b, S), as in one
of the specimens of the type (fig. 1, S). They are not, however, so short as in this case,
but, like the basal folds at their sides, reach the outer angles of the radial pentagon.

This is also the case in the other three varieties. In var. 4 the basal folds diverge
considerably at about the middle of their course (fig. 21, b.f), so that the dorsal inter-
radial furrow is here very wide, and then rapidly narrows towards its peripheral end.
In correspondence with this, the basal rays also widen somewhat from their narrow
central ends, and then begin to decrease in width as they approach the angles of the
radial pentagon (fig. 24, S) ; they are also marked by a slight median furrow along their
dorsal surface, as is the case in one specimen of the type (fig. 4, S). In varieties 2 and 3,
as in the type, and in var. 1, the basal rays are widest at their central ends (figs. 18,
19, 22, S). In both cases the basal grooves (figs. 17, 21, b.g) and basal folds (PL VII.

SECOND SERIES, — ZOOLOGY VOL. II. 11

100 ME. P. H. CARPENTER ON THE GENUS ACTINOMETEA.

fig. 4(1, b.f) are well developed and somewhat lancet-shaped in form, as in var. 1
(PI. YI. fig. 13, b.f, fig. 15, b.g).

In var. 2, and still more in var. 3, the mode of union of the hars of the basal bridge
(figs. 19, 22 b, b.b) with one another and with the basal rays (S), is seen very distinctly
at the central end of the latter; much more so than in Act.pectinata (PL V. fig. 8 b), in
which, as in the other specimens figured (PI. V. fig. 3 ; PI. VI. figs. 4, 11, 13), the
various elements of each compound basal are so completely united, that the lines of
unction between them become almost indistinguishable.

(§ 68) The complicated condition of the basals described above in Actinometra is not
altogether peculiar to this genus, as I was first inclined to believe ; for in Ant. Eschricldii
a basal star may be developed to a greater or less extent. In his paper on JPhanogcuia
Loven gives a diagram1 of the dorsal aspect of the pentagonal base of the calyx of this
species for comparison with that of Thanogenia. It shows five large rays extending
from the periphery of the rosette to the outer angles of the radial pentagon, with the
constituent elements of which they alternate in position ; and in the text he speaks of
them as belonging to the rosette. Unfortunately, his paper is written in Swedish, so
that I have been unable to ascertain precisely what his views were with respect to the
homologies of these rays.

His figure also shows five radial openings on the dorsal surface of the pentagonal
base, which correspond with the dorsal openings of the radial axial canals in Aid. rosacea
(PL IV. fig. 1G, Q).

In neither of the two specimens of Aid. Eschrichtii which I have been able to
examine do these openings exist, so that they are probably somewhat uncertain in their
occurrence, as in Ant. celtica (PL IV. figs. 3, 7) ; and in neither of these specimens is
the basal star developed to any thing like the extent that it is in the specimen figured by
Loven. In one (PL IV. fig. 10, S) the rays are excessively small and inconspicuous,
only extending for a very short distance along the dorsal interradial furrows (d.i.f),
while the corresponding basal grooves on the interradial elevations of the centrodorsal
piece are also very slightly developed (fig. 11, b.g). The basal bridge also, connecting
two successive rays of the basal star, is barely traceable around the inner margin of the
radial pentagon (fig. 10, b.b).

In the other specimen which I examined the basal rays were somewhat better
developed, occupying a larger portion of the dorsal interradial furrows, and extending
further outwards towards the margin of the radial pentagon ; although still remarkably
slender and delicate, somewhat as in Act. Solaris (PL V. fig. 3, S), and by no means so
large as in the specimen figured by Loven2.

The interior of the calyx of Ant. Eschrichtii is much simpler than that of Actinometra,

1 Loc. cit. p. 230, m.

2 Sinco the above lines were written, I have examined several other specimens of Ant. Eschricldii. None of
them have the radial openings above mentioned, nor arc the rays of the basal star so large as in the specimen figured
by Loven ; but they arc always present, though less regularly developed than in Actinometra. Further, ray work on
the 'Challenger' Comatulce has brought out the fact that a basal star is nearly always present in Antedon as well as
in Actinometra, so that the British species (Ant. rosacea, Ant. celtica) are remarkable for its absence, rather thun
Ant. Eschrichtii for its presence.

ME. P. H. CAEPEXTEE (XN THE GENUS ACTINOMETEA. 101

as Avill be seen by comparing PL IV. fig. 9 with PI. V. fig. 13, both of which represent
the internal aspect of two united first radials. In Actinometra (PI. V. fig. 13) there is
an abundant calcareous network (c.n) in connexion with the internal faces of the radials,
which are marked by well-developed radial and interradial furrows. In Ant. JEschrichtii,
however (PI. IV. fig. 9), the processes forming tins network are but little developed;
there is no axial interradial furrow, and even the radial one is indistinct, except near
the dorsal surface, where it passes between the raised edges of the two apertures
(x', y, y', z) of the central canal which unite with the inflected edges of a radial spout-
like process (p) of the rosette.

The interradial process of the rosette (o') is short and broad, but without the spout-
like character which it has in Actinometra — being simply directed, as in the normal
condition of Ant. rosacea (PI. IV. figs. 3, 7, 16), to the line of suture between the two
contiguous radials, to which it is attached between the two adjacent apertures (fig. 9, y, y')
of their central canals.

(§ G9) The remarkable variation in the extent to which the rays of the basal star may
be developed, as described above in Actinometra aud in Aniedon JEschrichtii, is due to the
fact that they are not calcifications in a nucleated protoplasmic network like the ordinary
elements of the skeleton. They are the result of a calcareous deposition, of a more
or less regular character, around the connective-tissue fibres which effect the synostosis
with the centrodorsal piece of every pair of contiguous radials along the line of
contact of the latter ; so that their position is, as we have already seen, interradial.

In PI. III. fig. 5 is seen the lower end of a vertical section cut transversely to the plane
of synostosis of two decalcified first radials (;•,) olAct. polymorpha, close to their peripheral
margin wdiere they are not concealed by the centrodorsal piece, so that the fibres of the
elastic ligaments connecting them with the second radials are cut somewhat obliquely (/J.

The threads of the protoplasmic network of which the organic basis of the radials is
composed pass somewhat rapidly at then surfaces into the connective-tissue fibrils (L)
which run horizontally between them and elfect the synostosis. These fibrils being very
closely set, the superficial portions of the calcareous reticulation forming the skeleton
of the radials, which are ossified around their ends, are very much more dense than the
central portion produced by calcification of the protoplasmic network.

PI. VIII. fig. 4 represents a section, from the same series as the previous one, across
the line of union of the same two radials (A, B), rather nearer to the centre of the calyx,
so that their dorsal surface appears no longer free, but partially covered by the centro-
dorsal piece (ccl). The lower portion of this section, more highly magnified, is seen in
PL III. fig. 6. The synostosis of the radial areas of the centrodorsal piece with the
dorsal surfaces of the first radials is effected by simple and not specially numerous con-
nective-tissue fibres (/) ; these pass directly from the protoplasmic basis of the one piece
into that of the other, in a direction vertical to the plane of the opposed surfaces, just
as in an ordinary synostosis. But in the direction of the interradii the course of the fibres
is different, and they have a deeper origin in the substance of the centrodorsal piece
than those which occupy the radial areas (PL III. fig. 6).

There are three principal masses of these longer inten adially placed fibres : — two smaller

14*

102 ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

ateral ones (S3), in which the fibres have the same direction as those occupying the radial
areas ; and a large median mass, in which the fibres ascend vertically for some distance
(S-,) and then diverge to the two sides (S{), where they pass into the protoplasmic basis
of the radials ; and the horizontal fibres (L) which pass between the radials fill up the
open angle caused by the divergence of the ascending fibres. There is thus a much greater
development of connective-tissue fibres, effecting the synostosis of the centrodorsal piece
with the radial pentagon, in the interradial than in the radial planes. This is well seen
in PI. VIII. fig. 3, which represents a longitudinal section through the calyx of Act.poly-
morpha. On the right side it is interradial, passing through the synostosis of the first
and second radials of the two radii, A, B ; and the connective-tissue fibres (S3) connecting
the centrodorsal piece with the edges of these two radii are longer and more abundant
than those on the left side (/), passing between the centrodorsal piece and the first radial
of radius B, which is cut longitudinally.

This is also seen in PI. VIII. figs. 5-8, which represent portions of four out of a series
of sections through a decalcified calyx of Act. pectinata. These are in the same plane as
the section of the calyx of Act, polymorpha represented in fig. 4, i.e. transverse to the synos-
tosis of the radii A, B on the one side of the centre, and to the radius B on the other.

.Pig. 5 represents a section, rather nearer the centre than fig. 4, passing vertically
along the axial interradial canal (a.i.c) ; beneath the dorsal end are seen the vertical
ascending fibres (S2), which have a much deeper origin in the substance of the centro-
dorsal piece of this species than in Act. polymorpha. The diverging fibres are not seen,
as they give rise by their calcification to the long basal ray (PI. V. fig. 8, S); and this
section passes through the depression at the central end of its ventral surface (PL V.
fig. 8 a, s) in which the axial interradial canal terminates.

Pig. 6 is somewhat nearer the centre, but still shows the long vertical fibres (S-,) in the
interradial plane, together with a portion of the central calcareous network aud the
axial radial canals (a.r.c) corresponding to the two radii A, B.

Pig. 7 is just beyond the centre, i. e. across the inner end of the first radial of B, so
that no vertical fibres arc visible, as they are only interradial in position. Two sets of
them, however, are seen in fig. 8, which shows the first radial of B cut transversely
rather further from the centre, so that the fibres (I) effecting its synostosis with the
adjacent radials of C and E are cut obliquely; beneath these are seen the interradial
ascending fibres (S2), which diverge slightly at their upper extremities (S\).

These diverging fibres and the upper ends of the vertical ones are the basis around
which the calcareous material forming the rays of the basal star is deposited. As the ven-
tral surface of the centrodorsal piece on which these rays rest is much higher at the centre
than at the circumference, it is impossible to obtain horizontal sections in which these
five rays are seen at all complete. Oblique sections, however, may be obtained in which
one or more of them are cut along the greater part of their length. Two such sections,
seen from their dorsal side, are represented in PL VIII. figs. 1 & 2 ; their lower left-hand
portions lie nearer the dorsal surface of the calyx than the upper right-hand portions.

The centre of fig. 1 is occupied by the fibrous envelope N of the quinquelocular organ,
from the dorsal portion of which cords (n.c) proceed to the cirrhi (cir). At the top and

MR. P. H. CARPENTER ON THE GENUS ACTINOMETRA. 103

right of the figure are seen portions of three first radials (t\), the central parts of which
are lighter than the more peripheral parts, as the section here passes through the slightly
developed unpigmented fibrous tissue (I) connecting the radials with the centrodorsal
piece (cd), the peripheral portion of which consists of the same pigmented protoplasmic
network as the substance of the radials.

Two of the synostoses between the latter are seen at the top of the figure ; but they
do not quite reach to the centre, where their place is occupied by the central ends of two
of the rays of the basal star (#1), the remaining three rays of which are visible in the
lower part of the figure for the greater portion of their length. Two of them are also seen
in the left or more dorsal portion of the next section (fig. 2), which also shows three
synostoses (L) between the radials in the right or more ventral part of the figure.

The sections of these rays of the basal star appear very dark, not from the presence
of pigment, which is entirely wanting in their fibrous basis, but because of the abund-
ance and very close approximation of the fibres of this basis ; these are the diverging
and vertical fibres seen in PI. VIII. figs. 3, 4, and more highly magnified in PI. III.
fig. 6 (Su S3). Just in the same way, in PI. VIII. fig. 3, the section appears much darker
on the right-hand side, where it cuts the closely approximated connective-tissue fibres
effecting the synostosis between two first radials transversely, than on the left-hand side,
where it passes through the more open protoplasmic network of the individual segment
of a single radial, I).

(§ 70) This great development of fibrous tissue along the interradial portions of the
centrodorsal piece and of the pentagonal base of the calyx accounts for the fact, often
mentioned already, that there is no pigment in the substance of the rays of the basal
star (which is a more or less complete calcification of the central portions of these inter-
radial fibrous masses), nor in the walls of the basal grooves on the centrodorsal piece, nor
in those of the dorsal interradial furrows on the inferior surface of the pentagonal base,
which are calcifications of the smaller lateral masses of long fibres running directly from
the organic basis of the centrodorsal piece into that of the first radials (PL III. fig. 6, 6'3).
These lateral fibres have a common point of origin in the substance of the centrodorsal
with the vertical and diverging fibres (#„ S2), around which the calcareous tissue of the
basal rays is deposited. It is therefore easy to understand that the calcification may
in some cases be so complete that the basal rays formed around the median fibres (S1} S2)
may become completely united with the walls of the basal grooves formed around the
lower ends of the two lateral fibrous masses (*$'.,) ; as is the case in the specimen of Act.
polymorpha represented in PL VI. fig. 8, where two of the rays of the basal star (S) are
so completely united with the floor and sides of the basal grooves in which they lie that
the line of junction between them becomes indistinguishable.

The fact that the rays of the basal star are calcifications in connective tissue and not
in the ordinary nuclear tissue which forms the organic basis of the other parts of the
skeleton, also affords an explanation of the great variations in the extent to which the
rays are developed. The general arrangement of the fibres constituting the interradial
portions of the synostosis between the centrodorsal piece and the radial pentagon is essen-
tially the same in Antedon as in Actinomctra. In Ant, rosacea and Ant. celtlca they

104 ME. P. H. CAEPEXTEE OX THE GENUS ACTINOMETEA.

never seem to undergo calcification, though this may take place in Antedon JEschrichti%
either only very slightly, as in the two specimens examined by myself (PI. IV. fig. 10, S),
or to a considerable extent, as in that figured by Loven. In Actinometra also the extent
of calcification of the rays of the basal star is very variable. In one specimen of Act.
polymorpha scarcely any trace of them is visible (PI. VI. fig. 11, S) ; in another they are
short, but otherwise well developed (fig. 4, S) ; while in others they may extend very
nearly to the outer angles of the radial pentagon (figs. 13, 18, 19, 22, 24, S).

In Act. pectmata, again, we have found them to be large and thick (PL V. fig. 8, S),
while in the closely allied Act. Solaris (PI. V. fig. 3, S) they were slender rods, only imper-
fectly calcified here and there — the intervening portions of the dorsal interradial furrow
seen in the prepared skeleton (d.i.f) being occupied in the fresh state by masses of fibrous
tissue, which are removed by the action of the alkali used in preparation.

(§ 71) We have seen that the basal circlet of Actinometra is somewhat complicated
in its nature, and consists of two entirely distinct elements, viz. a central rosette, which
we may fairly suppose to be the result of the metamorphosis of the embryonic basal
plates, as in Antedon rosacea, and five more or less completely ossified rays extending
from it in a peripheral direction.

The rosette is situated on the ventral side of the quinquelocular organ, from the fibrous
envelope of which proceed the primary basal cords. These are very short, and soon
bifurcate, so as to give rise to ten secondary cords, which pass through the ten apertures
(i\, v2, . . • zu 2a) m the peripheral portion of the rosette in order to reach the central
canals of the first radials.

As already remarked by Ludwig , we may fairly regard those elements of the skeleton
in which the bifurcation of the primary basal cords occurs as homologous throughout
the different genera of the Crinoids. This leads us to the conclusion that the rosette
of Antedon and Actinometra is homologous with the united central ends (at least) of the
basals of Pentacrmus, which are perforated by canals that lodge the five bifurcating
fibrillar cords proceeding from the dorsal angles of the quinquelocular organ, and not
from its ventral angles as in Comatula.

The question now arises, Where are we to seek for the homologues of the five rays of
the basal star in most Comatulce? Unfortunately the only type in which we find
a condition any thing like that described above is a fossil one, the Solanocrinus of
Goldfuss ; so that it is difficult to ascertain the precise relation of its basals to the canals
proceeding from the quinquelocular organ that was undoubtedly contained in the cavity
of its deep centrodorsal piece.

The upper surface of the latter, according to Goldfuss2, presents " five radiating
elevations on which the pelvis articulates." They correspond to the interradial elevations
on the ventral surface of the centrodorsal piece of Antedon and Actinometra.

The basals themselves vary in appearance in the different species. In S. costatus and
S. scroblcidatus they are only " fiinf schmale Strahlen die sich zwischen die Nahte der
Bippenglieder einsenken;" but in S. Jcegeri they are much wider, " so dass sie auf der
ganzen Gelenkfiaeke zusammenstossen und bier ftinf ausstrahlende Furchen zur Aufnahme

1 Beitriige &c. he. cit. p. 07. a Petrof. Germ. Joe. cit. p. L66.

ME. P. H. CAEPEXTEE OX THE GEXTJS ACTIXOMETEA. 105

der Saule bildeii " (p. 168). This species, however, is possibly not a Solanocrmus at all,
but the head of a stalked Crinoid.

In both these cases the peripheral ends of the basals appear on the external surface of
the calyx, between the centrodorsal piece and the radial pentagon, although the extent
to which they are visible is very different.

The long and narrow prismatic-shaped basals of S. costatus evidently represent
the five rays of the basal star of Actlnomctra. The iuterradial elevations on the
ventral surface of the centrodorsal piece are marked by five median grooves for the
reception of the basals, just like the basal grooves of Actinometra ; and these iuterradial
elevations are continued beyond the margins of the radial areas, just like the small
processes (t) in some species of Actinometra (PL VI. figs. 14, 15). They correspond with
five longitudinal ridges on the outer surface of the columnar centrodorsal which separate
the rows of cirrhus -sockets.

We do not, of course, know whether there was a rosette in S. cuslalns. I am
inclined to think that this was not the case, as the central ends of the five basals are in
: contact with one another laterally for a short distance, instead of being united by narrow
bars forming a basal bridge, as in Actinometra ; and their internal or proximal faces
were probably perforated by the opening of a short bifurcating canal lodging the fibrous
cords on their way to the central canals of the first radials, as in the closely similar basals
iof Pentacrinus asteria. Hence these five basals as a whole would represent the circlet of
compound basals in Actinometra, viz. the rosette together with the rays of the basal star.
Whether, however, only the united central ends of the basals of S. costatus represent the
embryonic basal plates, like the rosette of Antedon and Actinometra, or whether the whole
star results from a metamorphosis of the embryonic basals, is a question which must
remain in doubt, though the latter is by far the more probable supposition. Apart from
the analogy of Pentacrinus, as the peripheral ends of the basal rays extend beyond the
margin of the radial pentagon, it is hardly likely that they can be the result of cal-
cification in the iuterradial portions of the synostosis between the radial pentagon and
centrodorsal piece, as in Actinometra and Antedon Eschrichtii.

The calyx of the doubtful S. Jvgeri presents a great advance upon that of S. costatus
with respect to the development of the basals, which led Pictet L to propose the erection
of this species into a separate genus. Instead of being long and narrow, and in contact
only by their central ends, as in S. costatus and 5. scrobiculaius, they are broad and wedge-
shaped, and in contact along their whole sides, so as to form a complete calcareous disk
entirely separating the radial pentagon from the centrodorsal piece.

This is occasionally their position in Pentacrinus, though there are but few species of that
genus in which the basals are relatively so large and complete as in Solanocriuus Jcegeri.
In P. asteria, and in the two fossil species P. briareus and P. subangularis, they are
small and cuneiform and only in contact by their central ends, just as in S. costatus, so
that the greater portion of the radial pentagon is in contact with the top stem-seg-
ment. In P. Mulleri they are in contact for about half their length, and then diverge,
while in P. Wycille-Thomsoni they are completely united with one another along the

1 Op. cit. p. 2SS.

106 MR. P. H. CAEPENTEE ON THE GENUS ACTINOMETRA.

whole length of their sides, so as entirely to cut off the radial pentagon from the top
stem-segment, just as in S. Jcegeri.

There can therefore he little douht that the hasals of Pentacrinus are homologous with
those of Solanocrinus, and therefore analogous to the compound hasals of Actinometra,
which, as we have seen, are not entirely developed out of the embryonic basal plates.
It would seem, in fact, as if in Pentacrinus and Solanocrinus the embryonic basal
plates became directly transformed into the basals of the adult ; while in Comatula
they undergo metamorphosis into the central rosette by the absorption of the greater
portion of their dorsal or primary tissue, and the development of a secondary ossification
on the ventral side of the original plates.

In Ant. rosacea the metamorphosis is much more complete than in most Antedons
and in Actinometra, in which new skeletal elements are developed by a more or less
complete tertiary ossification in masses of connective tissue, that correspond precisely in
position, and to a certain extent also in shape, with the basals of Solanocrinus and Penta-
crinus. The latter being most probably direct products of the growth of the embryonic
basals are therefore strictly homologous only with the rosette of Actinometra, although
analogous in position to the whole circlet of compound basals in this genus, viz. to
the rosette and basal star taken together.

(§ 72) It is interesting to observe the different position of the basals with regard to
the chambered organ in Comatula and in the various species of the stalked Crinoids.

In Comatula this organ is situated in the cavity of the centrodorsal piece (PL VIII.
fig. 3) which is on the dorsal side, not only of the radial pentagon, but also of the rosette
or metamorphosed basals; and the nervous cords proceeding from its fibrillar envelope
to enter the central canals of the first radials come off from its ventral angles.

The large centrodorsal piece of Comatula is developed by the growth of the top stem-
segment of the Pentacrinoid larva. In Pentacrinus, which remains pedunculate
throughout life, the top stem-segment is the youngest and smallest. Its central
cavity is far too small to contain the quinquelocular organ forming the upper end of the
central axis of the stem, which contains five longitudinal chambers expanding slightly
at every nodal segment, where each of them gives off a single cirrhus-vessel1.

There is no special increase in the diameter of these chambers in the top stem-seg-
ments, and they do not expand into the large chambers of the quinquelocular organ
until near the level of the ventral surface of the basal circlet which surrounds the
dorsal half of the chambered organ. The ventral portion of the latter is contained in the
lower part of the central funnel-shaped space enclosed within the radial pentagon, where
it is surrounded by a very dense calcareous network, through which the axial prolongation
containing the superior continuations of the five chambers of the quinquelocular organ
ascends, on its way to enter the visceral mass, just as in Comatula. In consequence of
this relatively higher position of the chambered organ in Pentacrinus than in Comatula,
the nervous cords which enter the central canals of the first radials come off from its
dorsal angles, and not from the ventral ones as in Comatula.

Iu Comatula, therefore, the walls and floor of the cavity enclosing the chambered organ

1 Pentacrinus and Bhizocrinus, loe. eit. pp. 43-40.

ME, P. H. CAEPEXTEE ON THE GENUS ACTIXOMETEA. 107

are formed almost entirely by what was once a stem-segment ; while in Pentacrinus this
cavity is a part of the central space enclosed within the radial and basal pentagons, which
respectively form the ventral and dorsal portions of its side walls. Among the fossil
Apiocrinklce we find an intermediate condition between these two extremes. Thus in Ap.
mespiUformis, as seen in Goldfuss's figure 1, the first radials are small, but the basals are
very large and curved outwards, so as to enclose a large central cavity ; this we may
fairly suppose to have lodged a chambered organ, as Ludwig 2 has found that in Rhizo-
crmus, the modern representative of this family, the axis of the stem expands into a
chambered organ just as described above in Pentacrinus. I have found the same to be
the case in Bathijcrinus.

This organ is contained in H. 1 of of easts in a large and apparently simple segment,
described by Sars 3 as the expanded uppermost stem-segment. Pourtales4, however, for
reasons which will be discussed further on, regards it as composite and as representing
the five basals. In this case the relative position of the chambered organ is precisely
identical with that which we may suppose it to have occupied in Ap. mespiUformis,
namely, on the dorsal side of the radial pentagon, but not within the uppermost stem-
segment, as in Comidula. In Ap. rosacens5 the relative position of the chambered organ
must have been very much as in Pentacrinus, though slightly higher; for the cavity in
which it was contained was almost entirely enclosed between the enlarged first radials,
while the basals only form its floor and the very lowest portion of its side walls. Lastly,
in Ap. Ililleri6 the chambered organ must have lain altogether on the ventral side of the
basals; the superior surfaces of which form by their apposition the floor of a cavity whose
side walls are entirely composed of the adjacent inner faces of the contiguous first radials.

This condition is thus precisely the opposite of that which we find in Comatula, where
the cavity containing the chambered organ is not only altogether outside the radial
pentagon, but also on the dorsal side of the rosette or metamorphosed basals.

(§ 73) In the works both of Goldfuss and Miller may be found incidental suggestions
that the " basis," or circlet of basals, and, indeed, the whole of the lower part of the calyx
of the stalked Crinoids, may be regarded as representing expanded stem-segments, each
broken up into five parts. Muller, however, was the first to put this idea into a definite
form. He described the basals of Pentacrinus as a metamorphosed stem-segment 7, or as
" zerfallene Theile eines obersten Stengelgliedes ;" for they correspond in position with
the five leaf-like figures on the articular surfaces of the stem-segments which mark the
positions of the five longitudinal tendons.

The fibrous bundles composing these tendinous cords are separated from one another
by a very regular calcareous network, which is deposited around and between them,
somewhat as in the rays of the basal star of Actinometra. They are attached (by their

1 Petref. Germ., Taf. lvii. fig. 1, n.

' " Zur Anatomie des li. lofotensis," Zeitschr. f. wiss. Zool. Ed. xxix. p. 1:2:2.

3 Crinoides vivants, loc. tit. pp. -4, 12.

4 " On a new Species of Bhizocrinus from Barbadoes,'' Zoological Results of the Hassler Expedition, p. 28. Cam-
bridge, U. S., 1374. 5 Petref. Germ., Taf. lvi. fig. 3, e.

6 Petref. Germ., Taf. lvii. fig. 2, d. 7 " Bau des Pentacrinus" loc. eit. pp. 10, 25.

SECOND SERIES. — ZOOLOGY, VOL. II. 15

108 MR. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

upper ends) to the lower surfaces of the five basals in the same way as the ligaments of
the arms are attached to the brachial segments, the fibres of the one passing gradually
into the protoplasmic basis of the other. The basals cannot, however, be regarded as
simply ossifications in this fibrous tissue of the same nature as the basal rays of Actlno-
metra ; for, as shown above, there is every reason to believe that they are developed, like
the other elements of the skeleton, out of the embryonic basal plates ; although, so far as
position is concerned, they are precisely homologous with the calcareous deposits within
the tendinous cords of the stem.

On the other hand, the basal rays of Actinometra, which are similar in position, though
not in origin, to the peripheral portions of the basals of Pentacr'mus, are of the same
nature as the calcareous tissue of the leaf-like areas of the stem-segments, being simply
the result of the deposition of calcareous material around and between connective-tissue
fibres.

In many of the fossil Articulate Crinoids the lateral union of the basals with one
another is so very complete that the lines of junction between them are not always
visible, and the "basis" has therefore been described either as entirely absent or as
replaced by the uppermost stem-segment, which, according to Midler's view, it is sup-
posed to represent.

This is particularly the case in the Apiocrlnidce and in Eugeniacr'mus. Miller, who
was the first to describe the latter type l, mistook the first radials of E. caryophyllatus
for the basals, and described them as firmly anchylosed to what he supposed to be the
"superior columnar joint." Goldfuss2, however, rightly determined this last to be a
part of the first radials, which are very much prolonged downwards, while, at the same
time, he described the basals as replaced by the enlarged uppermost stem-segment,
which articulates with the inferior surface of the elongated first radials. E-cemer 3 did
not accept this view of Goldfuss's, although he recognized that the " superior columnar
joint " of Miller was simply a dorsal prolongation of the first radials ; but, like Miller,
he described these last as the basals.

It is most probable that Goldfuss's view is the truer one, as in Hagenow's figure 4 of
Eugeniacrinus Eagenoicii, in which the first radials are not prolonged downwards as in
E. caryophyllatus, the piece on which they rest, representing that which Goldfuss called
the enlarged uppermost stem-segment of E. caryophyllatus, is seen to be distinctly com-
posite ; for its external surface is marked by five sutural lines, alternating in position
with those between the first radials, and evidently indicating the lines of union of five
basals.

The "•eoloaical collection of the British Museum, which I have been able to examine,
thanks to the kindness of Mr. Henry Woodward, contains a very interesting series of
specimens from the Chalk which are labelled Apiocrinus elliptlcus.

In some of them the basals form a complete ring, separating the radials from the upper
stem-joint, which is very much enlarged. But in other specimens the basals appear

1 O+h cit. p. 111. - Petref. Germ. torn. cit. p. 102. 3 Lethoea Gcoguostioa, ii. Tbeil 4, p. 115.

4 Min. Jabrb. he. cit. ix. p. 13.

ATE. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA. 109

externally merely as small triangular pieces, not meeting laterally ; so that they exhibit
the same differences as the basals of Pentacrinus and of Solanocrinus. In many specimens
the sutures between the basals, radials, and top stem-joint are clear and distinct ; but in
others there is no trace of them at all, just as in some examples of JEugeniacrinus and
Bhizoci'inus ; but this is hardly a satisfactory reason for supposing the basals to be
internal and concealed, as has been done in the case of the last-named genus.

Even in some species of Pentacrinus the basals appear to be very closely united to
one another, and to assume the form of an uppermost stem-segment. Thus in P. sca-
laris, Goldfuss 1, there is no appearance whatever of small wedge-shaped basals, such as
are found in P. briareus and in P. asterla ; but, as remarked by the Messrs. Austin 2,
they appear to be united into a single plate, which resembles an " enlarged columnar
joint." The same was probably the case in the Jurassic genus Isocrinus, described by
Von Meyer 3, though it is, of course, possible that in both these cases the basals may
have been internal and concealed, as in Comatula.

I have endeavoured to show elsewhere4 that in the recent Phizocrimis we find a
strikingly similar case to that presented by JEugeniacrinus, viz., the sutures between the
basals, visible externally in one species and not in another, or, rather, not invariably in
another. In M. lofotcnsis the first radials rest upon a large and expanded apparently
simple segment, which was described by Sars5 as the expanded uppermost stem-segment;
and a small circular plate situated in the central vacuity between the first radials, with
which, as well as with the enlarged uppermost stem-segment, it is closely connected, was
regarded by him as representing the metamorphosed embryonic basals of Comatula.

Pourtales's observations 6, as well as my own subsequent ones, have led me to believe
that the piece called the enlarged uppermost stem-joint of P. loftensis by Sars and
Ludwig7 is composed (if not entirely, at any rate in great part) of five closely anchylosed
basals. Schluter s is evidently not acquainted with the evidence on which this view rests,
or he would scarcely suggest that B. Rawsonii might not be a Rhizocrinus at all, because
its basals differ from those described in P. loftensis by Sars and Ludwig ; although
these two observers are not themselves in accordance as to which parts of the interior
of the calyx are to be regarded as concealed basals.

Sir Wyville Thomson9 takes the same view as Pourtales and myself ; for he describes
how " in Rhizocrinus the funnel-shaped piece formed by the coalescence of the basals with
the fused first radials above and the dilated upper joint of the coalesced upper joints of
the stem beneath, makes up a large part of the cup ; " and his descriptions of the
calices of Hyocrinus and Bathycrinus, both genera allied to Rhizocrinus, together with

1 Petref. Germ. torn. cit. p. 173, Taf. Is. fig. 10, b.

2 'A Monograph on Kecent and Fossil Crinoidea,' p. 121 (Bristol, 1845).

3 " Isocrinus imd Chelocrinus," Museum Senkenbcrgianum, p. 251 (Frankfurt, 1837).

4 " Pentacrinus and Rhizocrinus," loe. cit. pp. 47-53. 5 Crinoides vivants, loc. cit. p. 4.

6 Hassler Expedition, Joe. cit. pp. 28, 29.

7 Rhizocrinus lofotensis, loc. cit. pp. 121, 122.

8 0/(. cit. p. 29. Schluter was unfortunately unable to make himself acquainted with Pourtales's memoir.

9 " Notice of new living Crinoids belonging to the Apiocrinidse," Journ. Linn. Soe. Zool. vol. xiii. p. 48.

15*

110 ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

the analogies of Apiocrinus and Eugenia crinus, strengthen Pourtales's view still more
completely.

The occasional fusion of the upper stem-joints with the lower part of the calyx, as
described ahove in the Apiocrinklce, is an excellent illustration of Midler's idea respecting
the correspondence between the basis and the stem-joints.

This correspondence, however, is by no means entitled to rank as a serial homology.
The earliest condition of the basals shows them to be five separate plates developed in a
spiral around the aboral ccelom of the Crinoid embryo1. They have distinct homologies
in the apical system of the other Echinoderms2; while the stem-segments, surrounding
the aboral coelom much in the same way as the basal circlet, arc simple undivided pieces
from the first, and seem to be almost or quite unrepresented in the other Echinoderms.

(vi.) The Second and Third Radlals.

(§ 74) The second radial oi Act. poly morpha (PI. VII. fig. 2), like that of Ant. rosacea,
is an oval, somewhat discoidal plate, having two nearly parallel faces — one internal or
proximal, articulating with the first radial, the other external or distal, articulating with
the third radial. The internal face (fig. 2 a) closely resembles the external face of the
first radial (PI. VII. fig. 1 b), with which it articulates, being divided transversely by a
large articular ridge (*) into a dorsal and a ventral portion ; the former is entirely occupied
by the fossa lodging the clastic ligament (/), which is particularly deep just below the
opening of the central canal (c.c). Prom the ventral margin of this opening arise
the two ridges which bound the intermuscular furrow (fx), and are joined near their
upper extremities by the transverse secondary ridges separating the large fossae (h)
that lodge the interarticular ligaments from those (f) lodging the flexor muscles of the
ray ; the latter are excavated in a pair of thin lamelke, which extend upwards from
the proper ventral margin of the plate, as is seen in a view of the distal face (fig. 2 b,
g). Besides the above-mentioned ridges and fossa?, which correspond to similar ones
on the distal face of the first radial, the proximal face of the second radial shows two
lateral processes, in which shallow fossa? (k) are excavated. These processes represent the
outer portions of the distal face, which is somewhat wider than the proximal one, as the
lateral faces are not set at right angles to the two terminal ones, but form an oblique
angle with the proximal face, so that the outline of the radial, when seen from the dorsal
or ventral side, is trapezoidal in form (fig. 2 c, d). The shallow fossae which are exca-
vated in these lateral faces lodge the ligamentous substance by which the second radials
are united with one another in pairs : the extent of this union is, as above remarked,
very variable in different specimens, being generally greatest where the number of arms
is largest (PL II. figs. 9, 11).

The external or distal face (fig. 2 b) is much simpler in character than the proximal
one, as no muscles are attached to the vertical lamella? which rise from its ventral
margin above the articular face proper. This last is divided by a vertical ridge (*) that

' Gotte, he. at. pp. 595, G20.

2 P. H. Carpenter, " On the Ural and Apical Systems of the Echinoderms." Quart. Journ. Alicr. Sci. xviii. (1878)
pp. 371, 382.

ME. P. H. CAEPEXTEE OX THE GENUS ACTING3IETEA. Ill

passes round the opening of the central canal (c.c) into a pair of lateral fossae (It), which
give attachment to the large interarticular ligament connecting the second with the third
radial. The proximal face is not quite vertical, but slightly inclined towards the distal
one, so that the ventral face is not much more than an edge. When the piece is viewed
from the ventral side, therefore (fig. 2 c), little else is visible but the fossa? for the
muscles (f) and interarticular ligaments (//) of the proximal face and the intermuscular
furrow (f) descending along its median line.

The second radials of var. 1 are very similar to those of the type, except that, as in
the first radials, the muscular fossae are relatively somewhat larger. In var. 2, however,
they are very much smaller (PL VII. fig. 5 a,/); and there are no vertical lamella; pro-
jecting from the ventral margin of the distal face (fig. 5 b), as is the case in the type.
The lateral fossae (k) lodging the ligamentous substance which connects the second radials
with one another are somewhat more marked, as the union of the second radials in pairs
is more complete than in the type, though not so complete as in varieties 1, 3, and 4.

The two latter also agree with var. 2 in the fact that the proximal and distal faces of
the second radials are nearly parallel, and less inclined to one another than in the type
and in var. 1 ; so that the fossae for the muscles and interarticular ligaments are barely
visible when the piece is seen from the ventral side (fig. 5 c), as there is a proper ventral
face. Its median line is occupied by a continuation of the furrow on the ventral surface
of the first radial (figs. 4 c, 5 c, v.r.f), while its lateral portions are divided up, in the
same way as those of the first radial, into secondary ridges and furrows.

(§ 75) The third or axillary radial of Act. polymorpha, which gives attachment to
two primary arms, presents three articular surfaces — an internal one corresponding to the
distal face of the second radial, and two external ones, inclined to one another, with which
the bases of the arms articulate.

The proximal face (PI. VII. fig. 3 a) is precisely similar in character to the distal face
of the second radial, being divided, like it, by a vertical ridge into two lateral fossae (//)
which lodge the interarticular ligaments. Its articular margin, when viewed from the dorsal
side (fig. 3 d), is perfectly straight, and does not project in the middle as in Ant. rosacea ;
so that the possible amount of lateral movement between the second and third radials
must be extremely slight. Two vertical lamellae (g) project from the uppermost margin
of the internal face ; but they do not form part of the surface of articulation with the
second radial, as they are excavated into fossae on their outer side for the attachment
of the proximal ends of the outer muscular bundles passing between the axillary radial
and the lowest segments of the primary arms. The two inner muscular bundles are
attached to the two sides of a projecting wedge-shaped process (cl) on the external or
distal face, the " clavicular" of Schultze1, which occupies the angle between the two ob-
liquely placed articular faces for the basal arm-segments. These are of precisely the
same character as the external faces of the first radials (fig. 3 b), consisting, besides
the muscular fossae (/) just mentioned, of two others for the interarticular ligaments (//),
and of a large dorsal fossa (j) lodging the elastic ligament, and separated from the other
two by a transverse articular ridge (i) , in the centre of which is the opening of the central
canal (c.c).

1 Loc. cit. p. 5.

112 ME. P. H. CARPENTER ON THE GENUS ACTINOMETEA.

The median line of the ventral face (fig. 3 c) is occupied hy a ventral radial furrow
continuous with that on the ventral face of the first radial ; it divides iuto two branches,
one of which passes on either side of the clavicular, in order to be continued on to the
basal arm-segments. The proximal end of this furrow is indicated by the deep notch
separating the two vertical lamella3 which project upwards from the proper internal
face (fig. 3 a, g), and through which the base of the clavicular is seen.

In the type and in var. 1 the lateral portions of the ventral face of the third radial
are plain, and not sculptured ; but in varieties 2-1 they are divided up by secondary
ridges and furrows (fig. 6 c), just like the ventral faces of the first and second radials
(figs. 1 c, 5 c). In these varieties also there are no vertical lamellae projecting upwards
from the ventral margin of the internal face (fig. 6 a), which is also the case in the ex-
ternal face of the second radial (PI. VII. fig. 5 b), as the muscular bundles passing between
the first and second radials, and between the third radials and basal arm-segments, are
smaller than in the type. In var. 2 there would appear to be more power of lateral
movement between the second and third radials than is the case in the type ; for
although, as in the type, there is no projection in the middle of the proximal articular
margin of the third radial, yet the distal articular margin of the second radial shows
a slight indication of such a median prominence (fig. 5 d), which is absent in the type.
It would seem though, to be replaced to a certain extent by the greater thickness of
the vertical articular ridge (*) around the opening of the central canal, which is seen, in
PI. VII. fig. 2 d, to project a little beyond the level of the dorsal surface of the radial ;
so that when the opposed ridges of the second and third radials are in contact with
each other, the third may possibly have a very slight power of lateral movement upon
the second, though by no means so great as in Ant. rosacea, in which the median pro-
minence on the internal articular margin of the third radial is very marked.

The second and third radials of var. 2 differ from those of the type of Act. polymorpha
and of all the other varieties in the very marked convexity of their dorsal surfaces, which
renders them considerably higher than the first radials ; so that when the whole calyx is
viewed from the exterior, the inner circle of first radials, which are only very little con-
cealed by the small centrodorsal piece, seems somewhat sunk within the outer circle
formed by the second and third radials.

This marked convexity is well seen in PI. VII. fig. 5, a, d, and fig. 6, b, d, especially
when these figures are compared with those of the corresponding parts in the type
(fig. 2, a, d, fig. 3, b, d).

List of Works referred to.

1. Adams, J., " Description of some Marine Animals found on the coast of Wales," Trans. Linn. Soc.
vol. v. p. 7 (London, 1800).

2. Agassiz, L., " Prodrome d'une Monographie des Radiaires ou Echinoderines," Annales d. Sci. Nat.
2mc se'r. Zool. vii. p. 288.

3. Agassiz, AL, " Revision of the Echini," Illust. Cat. Mus. oC Comp. Zool., No. vii. (Cambridge, U.S.
1872-71).

4. Austin, T. and T., Jun., 'A Monograph on Recent and Fossil Crinoidca' (Bristol, 1815).

ME. P, H. CARPENTER ON TIRE GENUS ACTINOMETRA. 113

5. Barrelieri, Jac, 'Plantrc per Galliarn, Hispauiam et Italiam observatae ' (Paris, 1714).

G. Beyrieh, E., " Ueber die Crinoideen des Muschelkalks," Abliaudl. d. k. Akad. zu Berlin, 1857.

7. de Blainville, H. M. D., 'Manuel d'Actinologie ' (Paris, 1834).

8. Bohlschc, AY., "Ueber Actinometra Bennettii und eine neue Coniatula-Art [Antedon Dubenii),"
Wiegm. Arcbiv fur Naturgesch. 1866, i. p. 92.

9. Broun, H. G., ' Die Klasseu und Ordnungen des Thierreichs,' Band ii. Aktinozoen (Leipzig,
1860) .

10. Carpenter, P. Herbert, "Remarks on the Anatomy of the Arms of the Crinoids, Part I.," Journ.
of Anat. aud Physiol, vol. x. April, 1876, pp. 571-585.

11. Carpenter, P. Herbert, The same, Part II., Journ. of Auat. and Physiol, vol. xi. Oct, 1876,
pp. 87-95.

12. Carpenter, W. B., " Researches ou the Structure, Physiology, and Development of Antedon rosa-
ceus, Part I.," Phil. Trans, vol. 156, pp. 671-756, jus. xxxi.-xlii.

13. Carpenter, W. B., "On the Structure, Physiology, and Development of Antedon rosaceus," Proc.
Roy. Soc, No. 160, 1876, pp. 211-231, pis. 8, 9.

14. Carpenter, W.B., " Supplemental Note to the above Paper," P. R. S., No. 169, 1876.

15. Columna, Fabius, ' Phytobasanus, sive Plantarurn aliquot Historia ' (Neapoli, 1592).

16. Dujardin and Dupe, ' Histoire Naturelle des Zoophytes Echinodermes' (Paris, 1862).

17. 'Encyclopedic Methodique,' " Histoire Naturelle des Vers, des Mollusques, des Coquillages et
Zoophytes" (Paris, 1789-1832), Partie des Vers.

18. De Freminvillc, "Memoire sur un uouveau Genre de Zoophytes de TOrdre des Radiaires," Nouv.
Bull. d. Sci. par la Soc. Philom. torn. ii. p. 349 (Paris, 1811).

19. Gegenbaur, C, ' Grundriss der verglcich. Anat/ (Leipzig, 1874).

20. Goldfuss, G. A., 'Petrefacta Germanise' (Dusseldorf, 1826-35).

21. Goldfuss, G. A., "Beitr. zur Petrefactenkunde," N. Acta Acad. Leop.-Carol. Nat.-Cur. xix. a.
1839, p. 318.

22. Goldfuss, G. A., "Beitr. zur Petrefactenkunde," Neues Jahrb. f. Mineral. (1841), p. 86.

23. Gotte, Al., " Yergleich. Entwickelungsgesch. der Comatula mediterranea," Archiv f. mikros. Anat.
Band xii. 1876, pp. 58 3-648, Taf. xxv.-xxviii.

24. Gray, J. E., " Notice on the Digestive Organs of the Genus Comatula, and on the Criiwidea of
Miller," Annals of Philosophy, N. S. vol. xii. p. 592 (1826).

25. Greeff, R., " Ueber den Bau der Echinodcrmen, 3. Mittheilung," Sitzungsb. d. Gesellsch. z. Be-
ford. der gesam. Naturwiss. zu Marburg, No. 11, 1872, pp. 155-169.

26. Greeff, R., The same, 4. Mitthciluug, " Ueber den Bau der Crinoideen," Marburg. Sitzungsb.
No. 1, 1876, pp. 16-29.

27. Greeff, R., The same, 5. Mittheilung, " Ueber das Herz der Crinoideen," Marburg. Sitzungsb.
No. 5, 1876, pp. 83-95.

28. Yon Hagenow, Fr., "Monographic der Riigen'schen Kreide-Versteinerungen, II. Abtheilung,
Radiarien und Annulaten," Neues Jahrb. fur Mineralogie, 1840, p. 664.

29. Heusinger, C. P., " Bemerkungen iiber den Verdauungskanal der Comatulen," Meckel's Archiv
f. Physiol. 1S26, p. 317.

30. Heusinger, C. P., " Anatomischc Untersuchungen der Comatula mediterranea," Zeitsch. f. organ.
Physik, Band iii. 1833, pp. 367-374, Taf. x., xi.

31. De Koninck, L., and H. Le Hon, ' Recherches sur les Crinoi'des du Terrain Carbonif ere de la Bel-
gique ' (Bruxclles, 1854).

32. Lamarck, J., ' Systeme d'Animaux saus Vertebrcs/ 2mo edit., Paris, 1816, torn. ii. p. 532.

33. Lange, W., " Beitr. z. Anat. und Histiologie der Asterien und Ophiuren," Morphol. Jahrb. Band
ii., 1876, pp- 211-286, Taf. xv.-xviii.

114 ME. P. H. CAEPENTEE ON THE GENES ACTINOMETRA.

34. Leach, W. E., ' Zoological Miscellanies/ vol. ii. p. 61 (London, 1815).

35. Leuckart, F. S., " Einiges iiber das Asteroiden-Geschlccht Comatula, Lam., iiberlianpt, und liber
C. mediterranea iusbesondere," Zeitsch. f. organ. Physik, Band iii. 1833, p. 385.

36. Linckii, Joliannis Henrici, Lipsiensis, ' De Stcllis Marinis liber singularis ' (Lipsiae, 1733).

37. Linnaeus, K., ' Systema Naturae/ editio decima tertia, pars vi. p. 3166 (Lipsiae, 1788).

38. Loven, S., " Phanogcnia, ett hittills okiindt slagte af fria Crinoideer," Ofvers. af Kongl. Ve-
tensk.-Akad. Forhandl. 1866, No. 9, pp. 223-233.

39. Ludwig, Hubert, "Beitr. zur Anat. der Crinoideen," Nachrichten von der Konigl. Gesellscb.
der Wisscnscb. und der G. A. Universit'at zu Gottingen, No. 5, 1876, pp. 105-114.

40. Ludwig, Hubert, ' Morpliologiscbe Studien an Echinodermen, I. Beitr. zur Anat. der Crinoideen'
(Leipzig, 1877) ; Separat-Abdruck aus der Zeitsch. f. wisscnscb. Zool. Band xxviii.

41. Ludwig, Hubert, "Zur Anatomie des Rhizocrinus lofotensis, M. Sars," Zeitschr. f. wiss. Zool.
Band xxix. pp. 101-130, Taf. v., vi.

42. Luidi, Eduardi, ' Lithophylacii Britanniei Iclmograpliia ' (Londini, 1699).

43. Luidi, Eduardi, ' Praelectio de Stcllis marinis Oceani Britanniei, nee non de Asteriarum, Entro-
chorum, et Encrinorum origine' (Oxford, 1703).

1 1. Meckel, J. F., " Ucbcr die Oeffnungen des Speisekanals bci den Comatulen," Meckel's Archiv f.
Physiol. Band iii. 1823, p. 470.

45. Metschnikoff, E., "Beitr. zur Entwickelungsgesch. einiger niederen Thiere/J Bull, de l'Acad. Imp.
des Sci. de St. Petersb. torn. xv. 1871, pp. 502-509.

46. Meyer, C. E. H., " Isocrinus und Ckelocrinus" Museum Scnkenbergianum (Frankfurt, 1837).

47. Miller, J. S., < A Natural History of the Crinoidea' (Bristol, 1821).

48. Miiller, J., " Ueber den Bau des Pentacrinus caput-Medusa," Abhandl. d. k. Akad. zu Berlin,
1843; Abst. inWiegm. Archiv f. Naturgesch. 1840, i.

49. Miiller, J., "Ueber die Gattungen und Arten der Comatulen," TViegm. Archiv f. Naturgesch. 1841, i.

50. Miiller, J., " Neue Beitr. zur Kenntniss der Arten der Comatulen/5 Wiegm. Archiv f. Naturgesch.

1843, i.

51. Miiller, J., " Nachtrag zu der Abhandlung iiber die Comatulen," Monatsb. d. Berlin. Akad. 1846,

p. 177.

52. Miiller, J., " Ueber die ^Gattung Comatula, Lam., und ihre Arten," Abhandl. d. k. Akad. z.

Berlin, 1849.

53. Miiller, J., " Ueber den Bau der Echinodermen," Abhandl. d. k. Akad. z. Berlin, 1853.

54. Norman, A. M., " On the Genera and Species of the British Echinodermata," Annals and Mag.
of Nat. Hist. ser. iii. vol. xv. p. 98.

55. d'Orbigny, A. D., ' Cours elementaire de Paleontologie et de Geologic stratigraphiquc ' (Paris,

1850-52).

56. Petiveri, .Tacobi, ' Gazophylacium Naturae et Artis' (Londini, 1711).

57. Petiveri, Jacobi, ' Aquatilium animalium Amboinensium Icones et Nomina' (Londini, 1713).

58. Philippi, R. A., " Alecto alticeps, n.^sp., eine [tertiare Comatula-Art von Palermo," Neues Jahrb.
f . Mineral. 18-14, p. 540.

59. Pictct, F. G., ' Traite de Paleontologie ' (Paris, 1857) .

60. de Pourtales L. F., " Contributions to the Fauna of the Gulf Stream at Great Depths," Bull,
of the Mus. of Comp. Zool. vol. i. no. 6.

61 de Pourtales L. F. " List of the Crinoids obtained on the coasts of Florida and Cuba by the
United States Coast Survey Gulf-Stream Expeditions, in 1867, 1868, 1869," Bull, of the Mus. of Comp.
Zool. vol. i. no. 11, pp. 355-358.

62 de Pourtales L. F., " On a new Species of Rhizocrinus from Barbadoes/' Zool. Results of the
Hassler Expedition, Must. Cat. of the Mus. of Comp. Zool. no. viii. pp. 27-31, pi. v.

ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETEA. 115

63. Retzius, A. J., ' Nova Acta/ Stockholm, 1783, no. 12, p. 234.

04. Eetzius, A. J., ' Dissertatio sistcns species cognitas Asteriarum' (Lundae, 1805).

65. Eceraer, F., In ' Lethcea Geognostica/ drittc Auflage, 1851, Theile iv. v.

66. Rosinus, M. R,, ' Tentaminis de Lithozois ac Lithophytis olim marinis, jam vcro subterraneis,
prodromus; sivc dc stellis marinis quondam, nunc fossilibus, disquisitio' (Hamburg, 1719).

67. Sars, M., ' Memoires pour servir a la connaissance des CriuoVdcs vivauts ' (Christiania, 1868).

68. von Scldottlieim, S., ' Nachtrage zur Petrefactenkunde,' Abtli. ii. (Gotha, 1823), p. 48.

69. Schultze, L., ' Monog. dcr Ecliinodcrmen des Eiflerkalkes ' (Wien, 1866). Besonders Abgedriickt
aus dem xxvi. Bande der Denkschr. der mathemat.-natu.rwissensc.haft. Classe d. kaiserl. Akad. d.
Wissensch.

70. Schweigger, A. P.,' Beobacht. auf naturhistorischen Reisen ' (Berlin, 1819).

71. Semper, C, " Kurze anatomische Bemerkungen iiber Comatula," Arbeiten aus dem zoolog.-
zootom. Institut zu TVurzburg, Band i. 1874, pp. 259-263.

72. Simroth, H., " Anatomic und Schizogonie der Ophiactis wrens, Sars, Theil I." Zeitsch. f. wiss.
Zool. Band xxvii. Heft 4.

73. Teuscher, R., " Beitr. zur Anatomie der Echinodermen, I. Comatula mediterranea" Jenaische
Zeitsch. f. Naturvrissensch., Band x. 1876, pp. 243-262, Taf. vii.

74. Teuscher, R., " Beitr. zur Anat. der Echinodermen, II. Ophiuridse," Jenais. Zeitsch. Band x. 1876,
pp. 263-280, Taf. viii.

75. Teuscher, R., "Beitr. zur Anat. der Echinodermen, III. Asteridre," Jenais. Zeitsch. Band x. 1876,
pp. 493-514, Taf. xviii., xix.

76. Thomson, Wyv., " On the Embryogeny of Antedon rosaceus, Linck {Comatula rosacea of
Lamarck)," Phil. Trans, vol. 155, 1865, pp. 513-544, pis. xxiii.-xxvii.

Addenda.

77. Baudelot, E., " Etudes generates sur le systeme ncrveux. Contribution a l'histoire du systeme
nerveux des Echinodermes," Arch, de Zool. experim. et g^nerale, tomei. pp. 177-210.

78. Carpenter, P. Herbert, "On some points in the Anatomy of Pentacrinus and Rhizocrinus," Joum.
of Anat. and Physiol, vol. xii. Oct. 1877, pp. 35-53.

79. Carpenter, P. Herbert, "On the Oral and Apical Systems of the Echinoderms, Part i.," Quart.
Journ. Micr. Sci. vol. xviii. new series, Oct. 1878, pp. 351-383.

80. Gegenbaur, Carl, ( Grundriss der vergleichenden Anatomic,' zweite Auflage (Leipzig, 1878).

81. Grube, Descriptions of three new Comatula (C. Icevissima, C. Mertensi, Act. borneensis) in Jah-
Iresber. d. schles. Gesellsch. 1875, Nat, Hist. Sect. pp. 54, 55*.

82. Ludwig, H., " Bcitrage zur Anatomie der Asteridcn," Zeitschr. f. wissensch. Zool. Band xxx.
pp. 150-212, Taf. v.-vii.

83. Schluter, Clemens, " Ueber cinige astylidc Crinoiden," Zeitschr. d. dcutsch. gcol. Gesellsch. Jahrg.
j 1878, pp. 28-66, Taf. i.-iv.

84. Thomson, Sir C. Wyvillc, "Notice of new living Crinoids belonging to the Apiocrinidae," Journ.
Linn. Soc, Zool. vol. xiii. pp. 47-55.

* I have, unfortunately, been unable to get a .sight of this paper, and only know of it from the reference to it in
Leuckart's ' Jahresbericht.'

SECOND SERIES. — ZOOLOGY, VOL. II.

10

116 MR. P. H. CARPENTER ON THE GENUS ACTINOMETRA.

DESCRIPTION OF THE PLATES.

The following letters denote the same parts throughout all the Plates.

A, B, C, D, E. The five Radii or Amhulacra.

Au A2, Bu B„, C„ Co, Du Z>2, Ev Et. The ten Primary Arms.

a.c. Axial ccelom. a.p. Axial prolongation.

a.i.c. Axial interradial canal. a.r.c. Axial radial canal.

a.i.f. Axial interradial furrow. a.r.f. Axial radial furrow.

An. Anal tube.

bu b2, b3, be, 0U). First, second, third, sixth, and tenth brachials. h.h. Basal bridge.

b.f. Basal fold. b.g. Basal groove.

b.mv Muscles between the radial axillary and the first brachials.

b.ni . Muscles between the second and third brachials.

c.c. Central canal of the calcareous segments of the rays and arms.

c.c'. Central canal of the eirrhus-segments. ch. Chambers of the quinquelocular organ.

rd. Centrodorsal piece. dr. Cirrhus.

cd.r. Centrodorsal coelom. el. Clavicular.

c.n. Calcareous network in the central vacuity of the pentagonal base of the calyx.

co.c. Commissural canals in the first radials. d.i.f. Dorsal interradial furrow.

cv.c. Circumvisceral coelom. d.r.f. Dorsal radial furrow.

dl} d.„ d.a. First, second, and axillary distichals. ep. Epithelial wall of the alimentary canal.

F. Central funnel-shaped space enclosed within the pentagonal base.

/. Muscular fossae. <]■ Vertical lamellae of the calcareous segments.

/j. Intermuscular furrow. gx. Their superior margins.

/2. Notch representing it in Ant. celtica. gr Their inner lateral margins.

gs. Ridge formed by the union of these in the first radials of Ant. celtica.

h. Fossae lodging the intcrarticular ligaments.

i. Transverse articular ridge. i.eo. Interradial commissure.

i.e. Interradial elevations on the centrodorsal piece.

ir.c. Intcrvisceral coelom. j. Fossre lodging the clastic ligaments.

k. Fossae lodging the ligamentous substance between the sides of the second radials.

L. Ligamentous substance between the sides of the first radials.

/. Ligamentous substance between the first radials and the centrodorsal piece.

L'. Incompletely decalcified portions of the skeleton.

Lr Interarticular and "1 ligaments between the first and second radials, or second and third

/,. Elastic J brachials.

L-,. Interarticular ligaments between the second and third radials. M. Mouth.

N. Fibrillar nervous envelope of the quinquelocular organ.

n. Axial nervous cords of the rays and arms. o. Interradial spout-like processes of the rosette.

n Their branches. o'. Interradial triangular processes.

v. r. Axial nervous cords of the cirrhi. o.b. Ovoid bodies. P. Peristome.

Pe. Uncalcified pcrisome between the radii. p. Radial processes of the rosette.

p. Small curved processes at the sides of a single basal.

px,pvp.a. First, second, and axillary palmars.

Q. Radial openings on the dorsal surface of the pentagonal base.

ME. P. H. CARPENTER OX THE GENUS ACTLXOMETEA. 11

Q'. Notches on the inner margins of the dorsal faces of the first radials.

q. Radial depressions on the centrodorsal piece. R. Rosette.

''ij rsi r-a- First, second, and axillary radials. r.ar. Radial areas on the centrodorsal piece.

r.c. Diverticulum from the radial ccelom into the substance of the first radial.

r.co. Intraradial commissure. r.o. Central opening of the rosette.

r.m. Muscles between the first and second radials. r.s. Radial space. S. Rays of the basal star.

S1} S& S3. Diverging, vertical, and lateral fibres effecting the interradial portions of the synostosis

between the centrodorsal piece and the radial pentagon.
s. Depressions at the central ends of the rays of the basal star.
tpi, sp.a. First and axillary suprapalmars. Sy. Syzygium.

t. Short processes at the angles of the centrodorsal piece.
U. Sockets for the attachment of the dorsal cirrhi.
u. Inner openings of the cirrhus-canals in the centrodorsal piece.
v.i.f. Ventral interradial furrow. v.r.f. Ventral radial furrow.

V, W, X, Y, Z. The five primary basal cords proceeding from the angles of the quinqueloeular organ.
Vi, V2, Wx, Wv Xu X2, Yx, Y2, Zu Zv The ten secondary basal cords produced by the bifurcation of

the primary ones.
i\, r2, ?i'„ W3, xv x2, yv yv glt Zv The apertures in the basals through which the secondary cords pass.
v, v', iv, 10, x, at, y, y', z, z . The corresponding apertures of the central canals on the internal faces

of the first radials.

In PI. II. figs. 9-11 indicate the position of the mouth (ventral) relative to the radial skeleton

(dorsal) .
* In PL V. fig. 4 indicates the passage of the ventral radial canal into the central calcareous network

within the radial pentagon by two openings, instead of by only one as usual.

Plate I.

Diagrams of the distribution of the ambulacra on the disks of different species of Comatula. The red

lines mark the interradial intervals. Figs. 1-4 copied from Muller.
Fig. 1. Antedon rosacea. Fig. 3. Act. Wahlberghii.

2. Actinometra solans. 4- Act. multiradiata.

In figs. 5-16 the tentaculiferous grooves are marked by dark lines, and the non-tentaeuliferous
grooves by fainter lines.

Fig. 5. Act. Solaris. Proportion of non-tentaculiferous arms, ^
6. Act. polymorpha : Type. » i's

7. „ „ » 18

„ <;

8. „ „ » 20

9. „ Var. 1. „ H

10. „ Type. „ 25

11. „ „ » ^6

12. „ „ » 28

13. ,, „ » 2S

14. „ Var. 2

15. „ Type.
1G. Var. 4. All the arms tentaculiferous

2 a

15. „ Type. » ;i

16*

118 ME. P. H. CAEPENTEE OX THE GENUS ACTINOMETEA.

Plate II.

Fig. 1. Diagram of the distribution of the ambulacra in a new Actlnometra from the Philippines.

2. Superior or ventral aspect of the disk of Act. polymorpha, var. 2, the oral pinnules having been

cut away near their bases, x 3.

3. Piece of an ordinary tentaculiferous arm of Act. polymorpha, from about the middle of its

length, seen from above, x 4.

4. Terminal portion of the same arm. x 4.

3. Piece of the; middle portion of a non-tentaculiferous arm, borne upon the same axillary as that

represented in figs. 3 & 4. Ventral view, x 4.
G. Termination of the same arm. x 4.

7. Distichal, palmar, and lower brachial segments of one of the radii of Act. polymorpha, var. 4,

showing the white line which occupies the middle of the dorsal surface of the skeleton, x 4,

8. CeDtrodorsal piece and one radius of a monstrous specimen of Act. polymorpha, showing the very

irregular form of the centrodorsal piece [cd] and the imperfect condition of one of the
palmar series, which consists simply of one axillary segment bearing brachials upon one of
its distal faces, and two suprapalmars [spusp.a) upon the other, x 4.

9. Diagram of the calyx of a small thirteen-armed specimen of Act. polymorpha : Type, x 4.

10. A similar diagram of a larger specimen with 2G arms. x 3.

11. A similar diagram of a specimen of var. 3, with 39 arms. x 2.

The * in these three figures ('.), 10, 11) indicates the position of the mouth on the ventral side of the
disk relatively to the radial skeleton of the dorsal side.

Plate III.

Fig. 1. Terminal comb of an oral pinnule of Act. pectinata. x 20.

2. Oral pinnule of Act. polymorpha, Type, x 10.

3. Oral pinnule of Act. polymorpha, Var. 4. x 10.

4. Portion of a horizontal section through the synostosis of two first radials of Pentacrinus Wyville-

Thomsoni. x 110.

5. Lower portion of a vertical section through the peripheral end of the synostosis of two first radials

of Act. polymorpha. x 110.

6. Lower portion of a similar section, taken rather nearer the centre of the radial pentagon, showing

the disposition of the fibres which effect the synostosis of the first radials with the centrodorsal
piece, both in the radial (/) and the interradial planes [Sit S2, S3). x 110.

7. Longitudinal section through one of the muscles (b.m._) and interarticular ligaments (L^ con-

necting the second and third brachials, x 110.
8-10. Cirrhi of Act. polymorpha. Type, and vara. 1 & 2.

8. Type, x G. a, adult; b, very young; c, nearly mature.

9. Isolated cirrhus-segments of the Type, x 40. a, a basal segment ; It, a terminal segment.

10. Cirrhus of var. 2. X 6.

11. Cirrhus of Tar. 1. x 6.

Plate IV.

Figs. 1-8 of Ant. celtica. All x 7.
Figs. 1 & 2. Centrodorsal piece, after removal of the cirrhi, as seen from its dorsal (fig. 1) and ventral
(fig. 2) sides.
3 & 4. Pentagonal bass of the calys, as seen from its dorsal (fig. 3) and ventral (fig. 4) sides.

ME. P. H. CAEPENTER ON THE GENUS ACTINOMETEA. 119

Fig. 5. Interior of the calyx, as seen after removal of the visceral mass.

6. Lateral view of the base of the calyx with the centrodorsal piece in situ.

7. Pentagonal base of the calyx of a smaller variety, as seen from its dorsal side.

8. Lateral view of the base of the calyx of the same variety, with the centrodorsal piece in situ.

Figs. 9-11 of Ant. Eschrichtii.

Fig. 9. Two united first radials, together with the portion of the rosette which is in connexion with
them, as seen from within, x 7.

10. Pentagonal base of the calyx as seen from its dorsal side after removal of the rosette occu-

pying its central cavity, x ■>\.

11. Centrodorsal piece seen from its ventral side. x 3£.

Figs. 12-17 of Ant. rosacea, all x 7, except fig. 13, which is x 15.

Fig. 12. Isolated first radial, a. ventral, /;. dorsal, c. internal aspect.

13. Abnormally developed rosette, with two spout-like interradial processes (o) and a basal bridge
(b.b.) connecting the ends of two of the radial processes (p). x 15. a. ventral, b. dorsal
aspect.
1 1. Lateral view of the base of the calyx with the centrodorsal piece in situ.
15. Centrodorsal piece seen from its venti'al side.
1G & 17. Pentagonal base of the calyx as seen from its dorsal (fig. 16) and ventral (fig. 17) sides.

Plate V.

The figures all x 7, except fig. S, which is x 15.

Figs. 1-4 of Act. Solaris.

Figs. 1 & 2. Centrodorsal piece as seen from its dorsal (fig. 1) and ventral (fig. 2) sides.

3 & 4. Pentagonal base of the calyx as seen from its dorsal (fig. 3) and ventral (fig. 4) sides.

Figs. 5-9 of Act. pectinata.

Fig. 5. Interior of the calyx as seen after removal of the visceral mass.

6 & 7. Centrodorsal piece as seen from its dorsal (fig. G) and ventral (fig. 7) sides.

8. An isolated compound basal, x 15. a. ventral, b. dorsal aspect.

9. An isolated first radial, a. ventral, b. dorsal, c. internal aspect.

Figs. 10-15 of Act. robusta.

Fig. 10. Internal aspect of an isolated first radial.

11-13. Two united first radials, together with those portions of the rosette which arc in con-
nexion with them, as seen from above (tig. 11), below (fig. 12), and within (fig. 13).
In these four figures (10-13) I. indicates a bristle passed along the axial radial canal; II. another
passed along the axial interradial canal; and III. a third, entering the central canal by one of the aper-
tures on the internal face (z1), and coming out through the aperture of the commissural canal (co.c.) on
the lateral face.
Figs. 14 & 15. Centrodorsal piece as seen from its dorsal (fig. 15) and ventral (fig. 14) sides.

Plate VI.
All these figures are x 7, except figs. 0, 18, 19, 22, which are all x 15.
Figs. 1-11 of Act. polymorpha, Type. Figs. 1-6, from one specimen.
Fig. 1. Lateral view of the base of the calyx, with the centrodorsal piece in situ.
2. The same parts seen from the dorsal side.

120 ME. P. H. CAEPENTEE ON THE GENUS ACTINOMETKA.

Fig. 3. Ventral aspect of the centre-dorsal piece.

4 & 5. Pentagonal base of the calyx as seen from its dorsal (fig. 4) and ventral (fig. 5) sides.
C. Two united compound basals as seen from their ventral side. x 15.
7, 8, & 9, from a second specimen.

7 & 8. Centrodorsal piece as seen from its dorsal (fig. 7) and ventral (fig. 8) sides. In fig. 8
three of the rays of the basal star are seen occupying the basal grooves (b. y), their proper
connexion with the rosette having been broken.

9. Pentagonal base of the calyx as seen from its dorsal side after removal of the rosette and

basal star.
10 & 11, from a third and abnormally developed specimen.

10. Centrodorsal piece seen from its ventral side.

11. Dorsal aspect of the pentagonal base of the calyx.

Figs. 12-15 of Act. polymorpha, var. 1.

Figs. 12 & 13. Pentagonal base of the calyx as seen from its ventral (fig. 12) and dorsal (fig. 13)
sides.
14 & 15. Centrodorsal piece as seen from its dorsal (fig. 14) and ventral (fig. 15) sides.

Figs. 16-19 of Act. polymorpha, var. 2.

Figs. 16 & 17. Centrodorsal piece as seen from its dorsal (fig. 16) and ventral (fig. 17) sides.

18 & 19. Two united compound basals as seen from their ventral (fig. 18) and dorsal (fig. 19)
sides.

Figs. 20-22 of Act. polymorpha, var. 3.

Figs. 20 & 21. Centrodorsal piece as seen from its dorsal (fig. 20) and ventral (fig. 21) sides.
22. An isolated compound basal as seen from its ventral (a) and dorsal (b) sides.

Figs. 23 & 24 of Act. polymorpha, var. 4.
Pentagonal base of the calyx as seen from its ventral (fig. 23) and dorsal (fig. 24) sides.

Plate VII.

In this Plate are shown the first, second, and third radials of the type of Act. polymorpha
(figs. 1-3) and of var. 2 (figs. 4-6).
The different aspects shown are designated as follows: — a. Internal or proximal face ; b. External
or distal face ; c. Ventral or superior face; d. Dorsal or inferior face.

Act. polymorpha, Type. Act. polymorpha, var. 2.

Fig. 1. First radial. Fig. 4. First radial.

2. Second radial. 5. Second radial.

3. Third or axillary radial. 6. Third or axillary radial.

Plate VIII.

All the figures are x 18. Figs. 1 & 2 of Act. polymorpha, Type.

Figs. 1 & 2. Two successive oblique sections through the base of a decalcified calyx, viewed from their
dorsal side.

Fig. 1 is the more inferior, i.e. nearer the dorsal surface. Its left-hand lower portion shows the centro-
dorsal piece only, with its marginal cirrhi (cir.) which receive fibrillar cords (n.c.) from the

ME. P. H. CARPENTER ON THE GENUS ACTINOMETRA. 121

central mass (A7) enveloping the quinqiielocular organ. Proceeding outwards from the
centre are seen five dark rays (Si), which represent the closely fibrillar organic basis of the
five rays of the basal star. In the upper part of the figure two of them are very short, only
their central ends being visible, as the section has here passed above the level of their outer
ends through the substance of three first radials (;•]).

Fig. 2. In this section only two of the basal rays (SJ are visible, as the greater part of it has passed
above the level of the synostosis (/) between the first radials and the centrodorsal piece. In
the centre are seen the chambers of the quinquelocular organ (c/i), with their ventral openings
into the vessels contained within the axial prolongation. At the right of the figure are seen
the lower ends of the axial canals, both radial (a:r.c.) and interradial (a.i.c.) ; their cavities
are generally crossed by transverse septa, which divide them up into two or three intercom-
municating smaller ones. The interval between every two of these canals is occupied by one
of the secondary basal cords (Y.2, Zu Z2, &c.), produced by the bifurcation of the short primary
cords (V, Y, &c.) proceeding from the angles of the quinquelocular organ. They are connected
with one another laterally by interradial and intraradial commissures (Leo. & r.co.) and
enter the central canals of the first radials (rx) in successive pairs, so that the axial nervous
cord (») of each radius is composed of fibres derived from two primary basal cords (Y„,Z1)
&c), just as in Antedon.

Fig. 3. A vertical longitudinal section through a decalcified calyx, passing on the right through the
synostosis of two first radials (A, B) and the fibrillar basis (<S2) of one of the basal rays, and
on the left through the segments (j-j, r2, r.a) of radius D. The first of these is united to
the centrodorsal piece by connective-tissue fibrils (/) similar to, but less abundant than,
those in the interradial portion of the section, around which, in the natural condition, the cal-
careous material forming one of the basal rays is deposited. The passage of the ventral
furrows (cr.f.,v.i.f.) into the axial canals (a.r.c, a.i.c.) is also well seen in this section.
Its centre is occupied by the quinquelocular organ, from the ventral portion of which the
axial prolongation (a.p.) rises into the circumvisceral ccelom (c.v.c), which, together with
the lower end of the wide axial ccelom (a.c), occupies the space between the ventral surface
of the skeleton and the lower or dorsal wall of the convoluted alimentary canal. It is
traversed by numerous connective-tissue septa, which divide it up into a system of spaces,
communicating freely with those both of the intervisceral and of the axial ccelom (ice, a.c).

Fig. 4. Transverse section through the synostosis of two first radials (A, B) near the peripheral
margin of the centrodorsal piece (cd.) , showing the radial (/) and the interradial (S„) fibres
which effect the synostosis between it and the united first radials. The latter form the
organic basis of one of the rays of the basal star.

Figs. 5-8. Four vertical sections, selected from a series, through a decalcified calyx of Act. pectinata.

5. Section through the adjacent inner ends of two first radials (A,B), showing the axial inter-
radial canal (a.i.c.) between them, and the open outer ends of the radial spaces (r.s.)
between their dorsal surfaces and the ventral surface of the centrodorsal piece. The central
ends of their axial nervous cords (//.) are cut very obliquely.

6. A section rather nearer the centre, showing the closed central ends of the radial spaces (r.s.)
of the same two radii (A, B) and their axial canals (a.r.c) ; also the four secondary basal
cords which unite in successive pairs (Z.2> V\ and V2, W-j) to form their axial cords, cut
obliquely.

7. A section from a little the other side of the centre, through the outer end of one of the
chambers (ch) of the quinquelocular organ, corresponding to radius D. The radial spaces
(r.s) of C and E arc cut almost longitudinally ; and above them, in the interior of the
radials, are seen the axial nervous cords, with one of the two secondary basal cords

122 ME. P. H. CAEPEXTEE ON THE GENUS ACTINOMETEA.

(Xu Y,) by which each is connected with the central nervous envelope of the quinquelocular
organ. The other branches (X2, Yj) of the two primary cords (X& Y) combine to form the
axial cord of the radius D. The inner end of its first radial is seen in the centre of the upper
part of the figure (r{), separated from those of G'&_Eby the axial interradial canals (a.i.c).
Fig. 8. A section rather further from the centre of the calyx, showing the first radial of D cut trans-
versely, with the closed central end of its radial space [r.s.) . At the sides of the latter are
the expanded dorsal ends of the axial interradial canals seen in fig. 7 ; they are received in
depressions (s) at the central ends of the rays of the basal star, which are ossified around
the vertical fibres (So) only, and not, like the stouter more peripheral portions of the rays,
arouud both vertical and diverging fibres, as is seen in Plate III. fig. 6, Su S2.

[ 123 ]

II. On some New Species of Nudibranchiate Mollmcafrom the Eastern Seas.
By Cuthbert Collingwood, M.A., M.B., F.Z.S., &c.

(Plates IX. & X.)

Read March 7th, 1878.

THE very considerable numbers of naked-gilled Mollnsca which have been found
upon our own shores would lead one to suppose that on other coasts, in which climatal
conditions were more favourable, they would be very commonly met with. Having
collected not fewer than twenty-eight species upon a very small section of our northerly

Wes'tof^r , H " T1'^1 SGarCh UP°U tr°piCal Sh°reS W°m ^ld - a^ant
haivest of these highly mterestmg and beautifully-tinted animals. In any such research

however, seasonal changes must not be overlooked. There can be no doubt in the mini

of any one who has ransacked tropical localities that in them the highest development

o colour and he most curious vagaries of form are to be found ; but I am persuaded

ha a zoologist who pays but a brief visit to a number of coasts in succession is far less

likely to make a considerable collection of species of Nudibranchiata than is anyone

who, confined to a single favourable locality for a considerable period, is thus able to

pursue h* examination of the same spot through all the different seasons of the year

Hence it is that the collections of Sir W. Elliot on the Madras coast (32 species), of

1 1. Kelaar on the Ceylon coast (42 species), and of Mr. Angas in New Smith Wales

hat'ofT^T S° laV§V^!fQ the V°^e °f thG 'Asteolabe' y^d but 18 new species,
that of the Samarang half a dozen species, and my own researches (the fruits of eve™
opportunity within my reach during a period of rather more than twelve months) resulted
only m the discovery of the 16 new species described in the following paper

The greater number of species of this group may be found upon stony shores near low-
water mark, and especially at low spring-tide. They adhere to the under side of stones

labW l ^ tlU'Uing tLem 0Ver" Alth°U-h SOme of them aPP<*r to be

capable of swimming, and most of them can float, branchiae downwards, upon the surface

ol the water I never saw them in this position in the sea, however calm. In fact theY

have invariably been found by me in a passive condition, like little shapeless masses of

oft, coloured matter, in depressions and crevices of stones, where they have probably

retired at that particular juncture, when the wash of low tide has disturbed the water

and rendered it both turbulent and turbid; from both of which disadvantages the same

spot would be free at all times, except at dead low-water. I have, however, dredged up

a very beautifully-eolourecl and delicate species from a depth of 20 fathoms, which

notwithstanding the rough handling of the dredge and the company of shells, corallines'

and sponges, seemed in no way incommoded or less lively. Again, some species of

JNudibranchiata, as Glaucus, appear to be free ocean-swimmers, and such I have taken

the towmg-net m the Formosa channel; while another species {Scyllcea pelagica) is

SECOND SERIES.— ZOOLOGY, VOL. II. 17

124 DR. CUTHBERT COLLING-WOOD ON SOME NEW SPECIES OF

abundant in the open ocean of the Atlantic, but always adherent by its clasping foot to
the fronds of the gulf-weed.

Nor do the exquisite colours of this group depend entirely upon geographical position.
On our own northerly coasts richly tinted species occur, such as Doris flammea, D. coc-
cmea, and many brilliant species of Eolis, &c. ; while Doris sordida is a Red-Sea species,
D. tristis is from the Madras coast, and D. exanthema from Ceylon ; and the same
remark also applies equally to other genera and families. Nevertheless, as a rule, the
more brightly coloured species are more commonly met with on tropical shores ; and of
the genus Ghromodoris, remarkable for their brilliancy, no species are found upon our
own shores, but, although some occur on the Mediterranean coasts, the majority are
characteristic of the hotter regions.

A very remarkable eiicunistance in the history of these delicate animals is their
extraordinary geographical distribution. Mr. Abraham remarks {I.e. jwstea) that the
well-known and almost first-described British species, Doris tuberculata, has also been
found in New Zealand on the one hand, and at Vancouver's Island on the other ; and
my own experience has remarkably confirmed the widespread habitats of certain species.
Interesting in this respect was the fact of my meeting with the same species, within a
few days' or weeks' interval, on different sides of the China Sea. Thus a Chromodoris
which I had already found in the Pescadores was the first thing I picked up, some time
later, on the reef of Labuan (Borneo). Veiy soon after I met with a second species at
Labuan, which I at once recognized as one I had already captured and figured in the
Haitan Straits, a little south of the river Min, coast of China. A Doris which I found
on a submerged reef in the centre of the China Sea, I afterwards found again on another
island off the coast of Borneo, west of Sarawak river. A minute species also (probably
an immature Trevelycma) I found on two occasions in localities separated by 150 miles
of sea.

A circumstance due, I think, to what I have already mentioned about seasonal changes
is worth recording. I had searched in vain upon some rocks in the harbour of Hong Kong,
and having mentioned this to a resident gentleman, who occasionally himself made the
same researches, he kindly offered to go with me and show me where he found them.
But we were equally unsuccessful, and could obtain none. The same thing occurred at
Labuan. Having showed my drawings to a gentleman interested in natural history, he
at once recognized them as animals of which he assured me there were many beautiful
species to be found ; and he also kindly conducted me to his hunting-ground, but equally
in vain, and to his surprise he could show me none where he had been in the habit of
finding them. In fact, although I met at Labuan with species I had already found
elsewhere, I only added one new species to my list at that place. In both these instances
I imagine I was at the localities mentioned at unfavourable seasons of the year, and not
at those seasons when my informants assured me they had met with numerous species.

I should mention, however, one other Labuan species which unfortunately I was
unable to record, owing to the following curious circumstance. It was a large tuber-
culated Doris, of a mottled grey colour, l< inches long, with capacious tentacles and
expansive gill-tufts, of which I found two or three specimens upon a reef in one of these

NEDIBEANCHIATE MOLLUSCA FEOM THE EASTEEN SEAS. 125

expeditions. I brought them home, and placed them in water till the next day, intending
by daylight to draw and describe them. But on visiting them in the morning, I found
that they had performed a spontaneous amputation of the mantle close to the body all
round. It was done as cleanly as if by a pair of scissors. A large Pyrula in the same
vessel was at first credited with this act ; but other specimens having been placed separate
in clean water, not being able to attend to them immediately, I found on visiting them
next day that they also had amputated their mantles and were destroyed. It appeared
to be a suicidal act, or " happy despatch," similar to the self-evisceration of Holothuriae
and breaking up of Comatula3 under the influence of the gradual fouling of the water.

These delicate and beautifully tinted animals are so entirely altered as to their form
and colour by spirit, that I can scarcely understand how new species can be satisfactorily
described from spirit-specimens. Alcohol bleaches their colours, and contracts to shape-
lessness the most beautiful elements of their form, the mantle, but more especially the
tentacles and branchiae, so that they bear no resemblance whatever to the living animal.
It keeps them, however, fit for dissection. The spicula, odontophores, buccal collars, and
other important classificatory characters are thus well preserved. Glycerine, while it keeps
for a considerable time their colour and form, renders them soft and comparatively useless
for dissection. The only way to retain a correct idea of their living character is to make
careful drawings of them in their active condition ; and whatever value the present
illustrations may have is founded upon this circumstance, for they are faithful to the life.

With regard to the two species of the Polybranchiate family of Phyllidiadae (the only
ones I met with, and found side by side in a rock-pool on the coast of Borneo), I have
consulted Bergh's elaborate paper in the ' Naturhistorisk Tidsskrift ' for 1868-69, and
also that in Heft 10 of Semper's ' Beisen im Archipel der Philippinen,' in both of which
papers species are described and figured ; but although there is some resemblance, I
cannot assure myself that the Phyllidiella pustulosa figured by him is the same as my
Phyllidia spcctabilis.

Nothing more need be added in these introductory remarks concerning the species to
be described in this paper. As to the works to which I have been indebted in the inves-
tigation, I would chiefly mention two, viz. the well-known and invaluable ' Monograph '
of Messrs. Alder and Hancock, published by the Bay Society, which has been to me as
a companion ever since its publication ; and, secondly, the " Bevision of the Anthobran-
chiate Nudi branchiate Mollusca," by Mr. P. S. Abraham, in the Proc. Zool. Soc. 1877 :
other memoirs are referred to in the text.

List of the Genera and Species enumerated in the present Paper.

Albania formosa.
Triopa Principis-Wallise.
Trevclyana felis.
Doridopsis arboresmis.
— rubra.
Phyllidia spectabilis.
Fryeria variabilis.
Scylhea pelagica.
Bornella nianuorata.

Doris pecten.

cresceutica.

Cbromodoris iris.

Bullockii.

aurco-purpurea.

tumulifcra.

tennis.

funcrea.

Alderi.

17*

126 DR. CUTHBERT COLLINGWOOD ON SOME NEW SPECIES OF

Order NUDIBBANCMIATA, Cuvier (1817).
Family DORIDIDiE, Alder and Hancock (1855).

Genus Doris, Linnaeus (1758).

Doris pecten, Coll., n. sp. (Plate IX. figs. 1-5.)

Length nearly -^ inch. Body oval, of a deep greenish-blue colour all over, studded
with minute papillae of a darker tint. Mantle large, covering the foot entirely. Dorsal
tentacles short, lamellated, dark-coloured, paler at the bases. Branchial, consisting of
seven or eight simple leaflets, are ranged in a crescentic form, the horns of the crescent
pointing forward and embracing a small crescentic area of a paler tint. Under surface
of the mantle paler than the upper, and spotted ; foot brownish below.

Two specimens of this little Boris were found in a rock-pool above low-water mark on
Bush Island, entrance to the harbour of Ke-lung, North Formosa, May 29. When at
rest, the posterior part of the mantle is drawn in, and the branchiae are seen in profile,
looking like a comb stuck in behind, whence its specific name.

Doris crescentica, Coll., n. sp. (Plate IX. figs. 0-8.)

Length 3 inches ; breadth 2 inches. Body broad, flat, tubcrculated. Mantle capa-
cious, covering the whole body, and largely projecting beyond the posterior extremity of
the foot ; very broad and round anteriorly, but about the middle of the body constricted
on either side to about one half its diameter ; edge very thin and flat, and puckered all
round the margin with numerous large and small folds. Plentifully covered with large
warty excrescences arranged crescentically and concentrically around the anterior and
posterior margins, where they are but slightly elevated : an elongated irregular excrescence
runs along the centre of the dorsum, commencing anterior to and between the tentacles,
and terminating at the branchiae,on cither side of which are arranged large irregular bosses
a quarter of an inch high. On either side of the branchiae there is a plain and thinner
irregular portion, larger on the left than on the right side. Dorsal tentacles large, cluh-
shaped, acuminated, and arising from a projecting eye-like sheath with irregular opening;
the club-shaped portions laminated, the peduncles smooth. Branchiae, of six compound
leaflets, much branched, and arranged in a wide, round, anal orifice ; the plumes nearly
equal in size. Mead with two small oral tentacles, concealed beneath the mantle.

Colour and general appearance. — General colour a brownish olive, the elevations and
bosses paler. Upon the large bosses is a tinge of pink, surmounted by a whitish apex ;
a similar pink tinge upon the thinner part of the mantle round the margin. The thin
non-tuberculated parts of the mantle on cither side of the branchiae straw-colour. Ten- |
tacle-sheaths same as the body-colour on which they arc situated, the tentacles themselves
somewhat darker. Branchiae pale brown, the stems darker, and the edges of the leaflets
whitish. Anal orifice Avhite. Under surface — anterior half of the mantle pale brown- i
pink; posterior half yellowish, a broad irregular reddish band immediately surrounding
the whole body.

NTJDIBEANCHIATE MOLLUSCA EEOM THE EASTEEN SEAS. 127

Of this remarkable species I found one specimen upon a coral block in a shallow patch
about 3 feet deep on the Fiery Cross Reef, China Sea, August 4th. The animal crawled
freely, but did not float on its back while under observation. When turned over on its
back it at once regained its natural position. The posterior portion of the mantle varied
in form from various spontaneous degrees of constriction ; sometimes it appeared to be
nearly of an oval form, at others it resembled a Tetlujs in shape.

Subsequently I obtained a specimen of this species on the island of Barundum west
coast of Borneo, of a gigantic size for Nudibranchiata, October 8th, upon coral blocks
between tide-marks. Of the two specimens found here, one was 6* inches Ion- and ^
wide. In all respects they resembled the one figured, even, to a great extent, in the
arrangement of the tubercles or bosses.

From a careful comparison of the description given by Kelaart* of his Boris exanthemata
with my drawings and descriptions of this species, I have come to the conclusion that
they are not identical, although evidently nearly allied. The form of the dorsal tentacles
the crescentic arrangement of the anterior tubercles, as well as several points in the
colour, to say nothing of the constriction of the mantle posteriorly (which might be
accidental), all point to different species; nor is the foot of the present species deeply
grooved and notched in front, as is the species described by Kelaart. I have named my
species from the crescentic arrangement of the tubercles on the anterior border of the
notseuni.

I am by no means certain, though, that the animal named Boris cerebralis, Gould f
may not be identical. The Nudibranch referred to, 5 inches long by 2f broad is stated
to have been taken from a reef in Sandalwood Bay, Feejee Islands. Comparing
Br. Couthouy's coloured sketch with my own, the two not being drawn exactly in a
similar position, considerable difference is manifest ; but if taken along with Dr Gould's
description in the text, it seems quite possible they may be the same species. Should
further research prove this to be the case, B. crescentica must necessarily be regarded as
a synonym.

Genus Chromodoeis, Alder & Hancock (1855).
Cheomodoris iris, Coll., n. sp. (Plate IX. figs. 9-11.)

Length 2J; inches. Body elongated, slightly tuberculated in profile, presenting two
rounded elevations, with depression between. Mantle scanty, disclosing the foot on either
side and posteriorly ; much waved at the margin. Anteriorly there is a thin-lobed fim-
briated veil, and posteriorly it is also divided into three lobes immediately behind the
branchial tufts. Borsal tentacles slender, conical, and finely lamellated, situated upon
slight elevations of the mantle. Bronchia composed of eight or ten simple leaflets
arranged in a cup-shaped form like the petals of a flower ; the leaflets delicately pinnate
Foot large, projecting about f inch behind the mantle, and tapering to a point posteriorly.

* Ann. & Mag. Nat. Hist, 3rd ser. vol. iii. p. 300 (1859).

t U.S. Exploring Expedition, Mollusca and Shells, Text, 1852, p. 208, and Atlas, 1856, pi. xxiii. figs 393 a-e
and Abraham, P. Z. S. 1877, p. 212. '

128 DR. CUTHBERT COLLINGWOOD ON SOME NEW SPECIES OF

Colour and general appearance. — Mantle deep blue, with a narrow edging of bright
yellow, and an irregular yellow stripe on either side of the median line, or broken up into
yellow spots. Large black roundish spots are scattered irregularly over the surface. The
foot is of a lighter blue colour, spotted irregularly with yellow and black spots ; the black
ones roundish, the yellow ones forming an irregular line. Branchiae and tentacles rich
vermilion, the latter arising from tricoloured bases.

Spawn, a spiral ribbon of a pale straw-colour, deposited under observation.

Two specimens of this splendid species I obtained in a basaltic rock-pool under a large
stone, at Makung, Pescadores Islands, May 12th, about midway between high- and
low-water marks. In August of the same year I obtained one specimen of the same
species from a reef on the shore of Labuan Island. Prom the beautiful combination
of the three primary colours presented in this species, I have given it the name of the
rainbow-goddess Iris.

Gould's Doris smaragdma* bears a distant resemblance, but is a smaller animal, more
greenish, with an indigo-blue margin and mantle, tentacles and branchiae yellow.

Under Ghromodoris rimcinataf and C. Semper 1 1 Dr. Rudolph Bergh has described
two new species from the Philippines ; but though in some points as to coloration and
markings there is an approach to C. iris, the distinction, in absence of dorsal and mar-
ginal yellow stripes, with greyish tentacles and branchiae and other particulars besides
geographical distribution, warrants specific separation.

Chromodobis Btjllockii, Coll., n. sp. (Plate IX. figs. 15-17.)

Length 2f inches. Body rather compressed, translucent. Mantle broad and square
in front, narrowing behind, and having the sides and posterior portion of the foot
uncovered. Dorsal tentacles ^o mcu l°ng> slender, consisting of a cylindrical pedicle,
smooth, and supporting a club-shaped, spirally laminated head. JBranchice consisting of
seven simple leaflets arranged in three sets, and arising from a thick retractile peduncle
situated in a cylindrical sheath of the mantle, a little more than halfway from the anterior
edge of the mantle to the posterior point of the foot. This peduncle gives off one leaflet
in front, and two lateral branches, each of which gives rise to three leaflets. The leaflets
are angular posteriorly, and edged with delicate papillae upon the anterior aspect.
Head concealed by the mantle, and bearing two small oral tentacles. Foot long and
fleshy, extending nearly three fourths of an inch beyond the posterior edge of the
mantle.

Colour and general appearance. — Body semitransparent. Head of a deep amethystine
tint, shading behind the dorsal tentacles through paler amethyst to reddish upon the
back ; an opaque white edging all round the mantle. Peduncles of the tentacles and
branchiae deep amethyst ; laminated portion of the tentacles and leaflets of the branchiae
deep orange-yellow. Foot pale amethyst, becoming deeper at the posterior portion, where
it is as deep as at the anterior part of the mantle.

* U. S. Exploring Expedition, vol. xii. p. 290 ; Atlas, pi. xxii. figs. 390 a-c.

t Semper's ' Reisen ini Arehipel der Philippine^' 1S77, Band ii. Heft 11, p. 479, and Hct't 10, pi. 53. figs. 5-12.

J Bergh, op. tit. supra, Exit 11, p. 482. pi. 55. figs. 2-7.

NUDIBEANCHJATE MOLLUSCA FEOM THE EASTERN SEAS. 129

Of this magnificent species I dredged one specimen in GO fathoms, off Recruit Island,
North Pacific, about 150 miles N.E. of Formosa. It was quite lively, and lived several
days, moving freely about, and floating foot uppermost, but never showing any tendency
to leave the surface of the water. I have named it after Captain Charles Bullock, R.N.,
a gentleman greatly interested in these studies, and from whom I met with much
assistance and courtesy.

Chromodoris aureo-purpurea, Coll., n. sp. (Plate IX. figs. 18-22.)

Length 1| inch. Bod;/ entirely covered by the mantle, except a small portion of pos-
terior end of the foot. Mantle ample, entire, smooth, broad anteriorly. In a second
specimen the edge of the mantle was somewhat indented in one or two places. Dorsal
tentacles small, club-shaped, upon cylindrical footstalk, curving backwards and outwards ;
the club-shaped portion finely laminated. Branch'ue of ten leaflets, forming a small
double ring ; the leaflets conical, pinnate, and diminishing in size from before backwards.
Foot slightly projecting behind mantle. Head rounded in front, with two minute oral
tentacles at the sides.

Colour and general appearance. — Upper surface with a general yellow tinge, and
covered over with small, irregular blotches of bright yellow, of a roundish or elliptical
form. Mantle edged with faint violet, and an irregular row of deep violet-shaded spots
running all round upon the faint edging, both being equally distinct upon the upper and
under side. The laminated portions of the tentacles dark violet, shading off at the foot-
stalk to the colour of the mantle. Branchiae, leaflets, and midribs deep violet at their
distal ends, becoming paler below, where they merge into the colour of the mantle.
Underside yellowish white, with a bright yellow spot invisible from above.

Two specimens were obtained under moderate-sized rough stones upon the inner shore
of Slut Island, Haitan Straits, on the Chinese coast, near low-water mark, June 30th.

These animals were very active, moving rapidly along with a gliding motion, at which
time their mantles were broad and flowing over the sides, the anterior portion somewhat
square. The tentacles were in constant and graceful motion, and they also frequently
floated, foot uppermost, on the surface. When at rest, however, the animal assumed a
nearly round form.

Mr. Andrew Garrett* has described and figured a new species (his Goniodoris Tryoni),
71 millims. long, from the Society Islands (Mus. Godeffroy, Hamburgh), which suggests
likeness to my Chromodoris aiirco-purpurea. His example is mentioned as creamy white,
margined with violet, and umber or tawny flesh-coloured branchial plumes and tentacles
tipped with violet. The body-spots, however, are deep black ocelli surrounded with
white. The branchial plume has 24 divisions, the posterior shorter than those in front,
and each terminally divided. Dr. R. Bergh later t refers to Garrett's species under the
designation Chromodoris Tryoni, and further describes and gives the anatomy of other
specimens collected by Prof. Semper, these differing slightly in tint and markings. The

* Troc. Acad. Nat. Sci. Philad. 1873, p. 232, pi. iv.

t In ' Ttoi.scn ira Archip. d. Philipp.' 1 S77, Band ii. Heft 11,1'. MO, and in Journ. d. Mus. Godeffroy, 1S77, Heft 14.
pi. iv. figs. 12-23.

130 DE. CUTHBEET COLLINGWOOD ON SOME NEW SPECIES OF

clearly defined black spots of Chromodoris Tryoni, as compared with the yellow blotches
in C. aureo-purpurea and other particulars noted by Bergh, give me reason to regard
that now described from the China coast as entitled to specific rank.

Chromodoris tumulifera, Coll., n. sp. (Plate IX. figs. 23-26.)

Length f inch. Body oblong, depressed, obtuse at either end. Mantle ample, covering
the whole body, except the posterior portion of the foot ; entire, tuberculated. Dorsal
tentacles smooth and delicate, club-shaped, the club-shaped extremities very finely lami-
nated, and twice as long as the cylindrical pedicle. Branchice small, consisting of nine
simple leaflets, the anterior largest, and diminishing in size posteriorly, the two hindmost
being rudimentary. Head crescentic, with two acute angles forming small tentacles,
one on either side.

Colour and general appearance. — Mantle translucent yellowish white, the upper surface
irregularly strewed with large, roundish, well-defined tubercles of a rich carmine colour
and tumuliform profile. Round the mantle runs a broadish band of chrome-yellow,
defined exteriorly, but somewhat fimbriated interiorly, leaving a narrow edging of the
mantle tint all round the outside : upon this edging are two carmine spots on the anterior
and two on the posterior angles of the mantle. The tentacles and branchial are of the
same tint as the mantle, the latter delicately formed and difficult of observation.

One specimen of this liandsome species found under a moderate-sized rough stone on
the south side of Slut Island, Haitan Straits, coast of China, in June. It was a somewhat
inactive animal, moving but slowly, but swimming occasionally foot uppermost on the
surface.

In August of the same year, being on the island of Labuan, on the very opposite side
of the China Sea, I met with this species more than once, at Pulo Pappan and Pulo Daat,
two islets between Labuan and the mainland of Borneo. These specimens were about
the same size as the Chinese one, but differed in that the carmine tubercles were more
numerous and encroached upon the chrome border ; the branchiae also were more deve-
loped, and I was led to imagine that the specimen figured was a young individual.

In colouring, this animal bears some resemblance to the Doris petechialis, Gould * ;
but this latter is 2 \ inches long, \\ inch broad, has vermilion-coloured tentacles, pinkish
branchiae, a more lemon-coloured margin and partially slate-coloured dorsum, and its
habitat is Honololu, Sandwich Islands. Dr. Gould admits his drawings are somewhat
imperfect, but sufficient for identification of the species.

The deep carmine spots recall the black ones of Garrett's Chromodoris Tryoni, I. c, but
in other respects the two cannot well be confounded.

Curomodoris tenuis. Coll., n. sp. (Plate IX. figs. 27-29.)

Length f inch. Body long and slender, very attenuated when in motion. Mantle
entire, covering the whole body, excepting the posterior portion of the foot ; broad and
squarish in front, and narrower from behind the tentacles backwards, bluntly pointed
posteriorly. Dorsal tentacles short and club-shaped, laminated, the suture anterior.

* U. S. Explor. Exped. Moll. vol. xii. p. 296 : Atlas, pi. 22. fig. 391.

NUDIBRANCHIATE MOLLUSCA FEOM THE EASTEEN SEAS. 131

Branchiae small, consisting of seven small and simple leaflets arranged in a circle, the
anterior leaflet somewhat larger than the others, and the posterior pair smallest. Foot
long and narrow, slightly tubular, projecting beyond the mantle posteriorly.

Colour and general appearance. — Mantle opaque white with a slight tinge of yellow,
especially on the anterior portion, edged with chrome-yellow, slightly shading off inte-
riorly. The whole mantle is covered with minute roundish spots of carmine, irregularly
distributed, absent only from the most anterior portion, the spots varying in size from
mere specks to roundish definite spots. Tentacles yellowish, but not so bright as the
border of the mantle; the bases whitish. Branchiae pale yellow. Foot edged with
chrome posteriorly, like the mantle. Under surface yellowish, foot and mantle with a
faint edging of chrome-yellow, the carmine spots showing through at the sides of the
head.

Two specimens were found on the under surface of a block of coral in a shoal patch of
reef in the midst of the China Sea, named Eiery Cross Reef. It is possible the spot
might be uncovered occasionally at low spring tides, but was now 3 feet under water.
Notwithstanding this, however, tbese little creatures when captured were continually
crawling out of the water and resting upon the dry edge of the vessel in which they were
contained, under which circumstances they had a short and stumpy aspect. When placed
in the water they were very lively and at once commenced crawling, having first stretched
themselves to double their previous length, with a proportionate tenuity. While crawling
they had a remarkably slender and linear appearance. They also swam freely on the
surface, foot uppermost. (August.)

In most respects the Chromodoris tenuis agrees with the Doris aspersa, Gould*, save
size, his specimen being 1^ inch long by \ an inch broad, and its habitat Vincennes Island,
Paumotu group. It may be questionable in this case whether we have specific distinction,
or whether size and other slight variation may not be attributable to difference of age,
sex, or geographical range. The D. aspersa, D. cerebralis, &c, given by Gould, were
drawn from nature by Mr. Joseph P. Couthouy, Naturalist to the U. S. Explor. Exped.,
I and therefore may be deemed reliably correct.

• Chromodoris eunerea, Coll., n. sp. (Plate IX. figs. 30-33.)

Length If inch. Body simple, stout, except when actively in motion, when it becomes
I attenuated, obtuse in front. Mantle smooth and entire, having the posterior portion of
i the foot exposed during progression, well rounded anteriorly. Dorsal tentacles small,
i arising from a small simple sheath, laminated, with scarcely any pedicle. Branchiae
composed of seventeen or eighteen leaflets arranged in a convoluted form, the larger
i leaflets in front, the smaller behind, leaflet irregularly branched. Head concealed by
the mantle and supporting a pair of oral tentacles.

Colour and general appearance. — Border of mantle narrowly edged with orange;

general aspect of the upper surface a rich dark brown, with yellowish-white or white

lines, following the direction of the border of the mantle, but in some places slightly

ramifying and sometimes anastomosing. The posterior and exposed portion of the foot

* U. S. Explor. Exped. Mollusca and Shells, p. 304 ; and Atlas, pi. 25. figs. 399 a-c.

SECOND SERIES. — ZOOLOGY, VOL. II. 18

132 DE. CUTHBERT COLLLNGWOOD ON SOME NEW SPECIES OF

is bordered and streaked like the mantle. The tentacles have a white ring, sometimes
two rings, round the sheaths ; they are reddish above, deepening to dark brown below ;
rows of opaque white spots are arranged nearly regularly, parallel with the lamina? ; at
the apex is a small whitish ring. The branchiae are reddish, the inner side of the stems
of the leaflets marbled below with brown and white ; the whole branchial tuft studded
like the tentacles with minute spots of opaque white, having a very beautiful appearance.
The under surface of the foot is white, with an orange border, like the mantle ; the sides
of the foot striped as above, only the white lines are whiter, but less distinct, and the dark
body-colour paler.

This is a remarkably handsome species. The longitudinal white striae become widened
when the animal is at rest, and very much attenuated when in progression. The branchiae
and tentacles look as though studded with little pearls. It is somewhat remarkable
that these animals never floated upon the surface of the water while under observation,
and if placed upon their backs always immediately turned over.

These specimens wrere taken upon a reef east of the island of Labuan, Borneo, and one
on the adjacent islet, Pulo Pappan, in August.

Chbomodoris Alderi, Coll., n. sp. (Plate IX. figs. 34-37.)

Length 2 inches. Mantle capacious, covering the head, squarish in front, slightly
emarginated. Dorsal tentacles very small, and placed upon very short pedicles, finely
laminated, having the commissure anterior ; they have precisely the appearance of small
cochineal insects. Branchial of ten simple four-sided angular leaflets, some of them
bifurcating near the apex, arranged in an imperfect circle, curving outwards and sur-
rounding the anal orifice. Head with two small white oral tentacles. Foot somewhat
tubular posteriorly, and extending about \ inch beyond the mantle.

Colour and general appearance. — General body-colour an opaque yellowish white or
cream-colour, a border of bright orange running all round the edge of the mantle and
projecting portion of the foot. Back beautifully marbled with reddish brown, an irregular
row of carmine spots placed all round the marbled portion, between it and the orange
border. Tentacles laminated alternately with crimson and white. Branchiae reddish,
the angles crimson. Under surface of a delicate transparent white.

A most beautiful species, slow in its movements, which were confined to crawling and
floating upon its back. One specimen only, found between tide-marks in a sandstone
rock-basin in Ke-lung Harbour, North Formosa, May 31st. I have named it after the
late Mr. Joshua Alder, whose name will always be connected with the history of this
elegant order of Mollusca.

Genus Albania, Collingwood, nov. gen. (1878).

Corpus depressum, niolle, semipellucidum. Notseuin amplissimum, undulatum et inversum. Rkino-
plioria flexibilia, sine vaginulis. Caput velo bilobato obsituru. Branchiae e circa 7 foliolis,
separation retractilibus, compositse.

Body depressed, soft, semitransparent. Notaeurn (mantle) ample, undulated, and
turned up at the sides. Dorsal tentacles (rhinophoria) flexible, without sheaths.

NTTDIBEANCHIATE MOLLUSCA FEOM THE EASTERN SEAS. 133

Branchiae consisting of about seven compound leaflets, each separately retractile. Head
with a bilobated crenated veil.

I have named this genus after the late If r. Albany Hancock, a gentleman so well
known in connexion with the history of these animals. For euphony's sake I have so
adapted the name as to make it agree with Formosa, the native island of this elegant
species. The genus must be placed near BZexabranchus.

Albania Formosa, Coll., n. sp. (Plate X. figs. 1-5.)

Length 2 inches. Body extremely delicate, almost semitransparent. Mantle broad
and capacious, forming, as it were, wings or fins on either side ; the edges turned over the
back when at rest. Dorsal tentacles large and rather thick, consisting of a conical,
bluntly pointed, finely laminated portion, with the commissure anterior, mounted upon
a cylindrical pedicle of equal length, without a sheath, non-retractile. Branchial con-
sisting of seven compound leaflets, each having three or four or five branches, and arising
from a common thick pedicle, but separately retractile, the whole forming a ring near the
posterior extremity of the mantle ; very delicate and almost transparent. Head with a
bicrcscentic veil, and studded round with a fringe of minute papilla?. Foot extending
j inch beyond the mantle posteriorly, somewhat rounded.

Colour and general appearance. — General tint a pale rose, darker and richer on the
back, forming a Vandyke pattern nearly regular on either side. Edge of the mantle
opaque white, with a wide inner border of crimson, the whole studded with minute
whitish translucent points, the greater or lesser abundance of which effects the gradation
of colour. Laminated portion of the tentacles crimson ; pedicles pale rose. Fringe of
the veil orange ; veil and posterior portion of the foot yellowish ; under surface pale rose.

One specimen of this singular and beautiful Nudibranch was taken in a red sandstone
rock-pool in Ke-lung Harbour, North Formosa, May 31st. When placed in a vessel of
sea-water it commenced swimming freely with a vertical vermicular movement and
extreme grace. The mantle was spread out wide, the tentacles thrown back, like ears,
ind the anterior and posterior extremities of the body thrown upward till they met above,
lien partially thrown back, the mantle waving in a vermicular manner from anterior to
posterior edge. It continued swimming like this for a considerable time. It did not
:rawl about like other Xudibranchs, but when not swimming remained in a more or less
contracted form, the mantle constantly changing its aspect. When, however, I turned
It over, it floated on its back like its congeners. I have called the species Jbrmosa, both
I roni its beauty and the island of which it is a native.

Family POLYCERID^], Alder & Hancock (1855).
Genus Triopa, Johnston (1838).

tsiopA Princlpis-Walll^, Coll., n. sp. (Plate X. figs. 6-11.)

Length f inch. Body slender, narrow, rounded in front, obtusely pointed behind.
'lantle scanty, just covering the body, smooth, and furnished with papilla; round the
nterior portion and along the sides. The anterior papilla?, eight in number, arranged

18*

134. DE. CUTHBEBT COLLINGWOOD ON SOME NEW SPECIES OE

in a crescentic form as a veil ; the lateral papillae five on either side. These papillae
consist each of a conical stem with pinnae, the lateral papillae being about twice the
length of those round the head. Dorsal tentacles club-shaped, the upper half swollen
and finely laminated, upon a cylindrical pedicle. Branch he of five leaflets arranged
round the anal orifice upon an elevated portion of the body ; the anterior leaflet much
the largest, and the two posterior minute. Each leaflet simply pinnatifid, and feather-
like in general aspect.

Colour and general appearance. — Body of a general pale orange-yellow, darker between
the tentacles and along the median line, spotted irregularly with minute dots of orange.
Upper half of the papillae with larger spots of orange. Pinnae translucent yellowish.
Branchiae pale orange-yellow. Tentacles spotted with orange about the central parts of
the laminated portion.

One specimen found beneath stones near low-water mark, on Slut Island, Uaitan Straits,
coast of China, June 30th. It was rather sluggish in its habits, but swam on the surface,
foot uppermost. I have named the species from the resemblance of the branchiae to the
well-known crest of the Prince of Wales.

Genus Trevelyana, Kelaart (1858)*.
Trevelyana felis, Coll., n. sp. (Plate X. figs. 12-14, immature.)

Length % inch. Body simple, smooth, of a uniform scarlet colour, the intestines
showing darker upon the dorsal surface. Mantle indistinct, covering the entire body.
Dorsal tentacles two, perfectly simple, conical, scarlet. Branchice, none visible.

Several specimens of this little animal occurred upon a stone in a tide-pool on the
basaltic shore of Makung, Island of Pong-hou, Pescadores, in May ; and in June, six
weeks later, I also found it on Slut Island, Haitan Straits. It was very active and flexible,
assuming at different times the most singular forms, resembling in turn a fox, a rabbit,
a cat, according to its different attitudes. It swims like the other nudibranchs on the
surface of the water, foot uppermost. I cannot speak with great certainty of it, but
believe it to be an immature species of Trevelyana, the trivial name being added to call
other observers' attention to it. Qu. In what respects does it stand to the T. (Stenodoris)
rubra, Peascf , from the Pacific ?

Family DOBIDOPSID^E, Alder and Hancock (1864).

Genus Doridopsis $, Alder and Hancock (1864).

Doridopsis arborescens, Coll., n. sp. (Plate X. figs. 15-17.)

Length If inch. Mantle capacious, enveloping the body, and overlapping the foot
laterally and anteriorly; the edges deeply cut and puckered all round; smooth and

* Trevelyana, Kel. Journ. Asiatic Soc, and Aim. & Mag. Nat. Hist. 3rd ser. (1858) vol. i. p. 257. Messrs. Alder
and Hancock, Trans. Zool. Soc. vol. v. p. 132 (footnote) suggest that "this genus may possibly bo synonymous with
the Cfymnodoris of Stimpson, Proceed. Philad. Acad. Nat. Sc. 1855."

t Amer. Journ. Conch, vol. ii. p. 206, pi. 4. fig. 2.

J See " Indian Nudibranchiato Mollusca," Trans. Zool. Soc. vol. v. p. 124.

XUDIBRANCHIATE MOLLTJSCA FROM THE EASTERN SEAS. 135

velvety. Dorsal tentacles rather large, consisting of a thickish footstalk pointing
forwards and a laminated club-shaped portion curving backwards; the whole seated
within the rim of a narrow retractile sheath. Branchiae very large, consisting of about
seven compound ramified leaflets, radiating from around the anal aperture, and situated
near the posterior edge of the mantle. The anterior leaflets (pointing forward), when
fully expanded, reach nearly halfway over the back; the posterior leaflets smaller. Each
leaflet finely pinnatifid, and the whole forming a beautiful star nearly an inch in dia-
meter, concealing the posterior portion of the animal. Foot large, occasionally visible
beyond the posterior edge of the mantle ; deeply cleft posteriorly.

Colour and general appearance. — The whole of the mantle a rich deep velvety blackish
brown, edged with fight chestnut. The peduncles of the tentacles blackish, translucent ;
the laminated portion like the mantle, and tipped with chestnut ; branchial leaflets
dark brown, tipped with light greyish at the edges. Underside of the mantle blackish,
translucent ; and foot light brown, shading to chestnut along the margin.

Two specimens of this handsome Nudibranch were found among rough stones on Slut
Island, Haitan Straits, coast of China, near low-water mark, June 30th.

The animals were sluggish, not moving much, nor fast; they floated readily, foot
uppermost, on the surface. The mantle was so capacious that in some attitudes they
appeared as broad as long; but when ordinarily crawling, the large and beautiful branchiae,
were very conspicuous, occupying nearly the posterior half of the body, and concealing
the posterior margin of the mantle and end of the foot, which at that time projects
beyond the mantle.

Messrs. Alder and Hancock * have shown that in the case of Doridopsis nigra there are at
least three or more varieties with gradation of tints and markings, and inhabiting the
coasts of the Loochoo Islands, Madras, and Ceylon. Our species, D. arborescens, appears
quite distinct ; but nevertheless, with such tendency to variation in a form closely related,
careful comparison with a series might cause a different view to be taken as to its
separation or identity with species already described by other authors.

Doridopsis rubra, Kelaart. (Plate X. fig. 18.)

Length 1| inch ; breadth 1 inch. Body thick, sluggish, opalescent. Mantle capacious,
thin, covering the whole body, except the posterior portion of the foot ; smooth, semi-
transparent, folded posteriorly during progression. Dorsal tentacles pyramidal, short
and thick, curved laterally, the footstalks smooth, as long as the upper portions, which
are swelled, pointed at the extremities, and laminated. Branchiae consisting of six
compound leaflets, like feathers, somewhat unequal in size, surrounding the anus, retrac-
tile. Read small, covered with the mantle. Foot left uncovered posteriorly during
progression.

Colour and general appearance. — Mantle of a rich rose-colour, darker in the thicker
parts of the back, and paler upon the thinner portions of the sides, marbled on the back
Avhen contracted, from corrugation of the surface. Branchiae of the same rose-tint as

* Ind. Nudib. Moll., Trans. Zool. Soc. vol. v. p. 128.

136 DE. CUTHBEET COLLINGWOOD OX SOME NEW SPECIES OF

the mantle. Dorsal tentacles of a rich rose-colour, the footstalks paler. Underside of
mantle pale rose, and underside of the foot inclining- to yellowish.

One specimen, found under a stone hetween tide-marks in Singapore harbour, immedi-
ately west of the town, in December. It was sluggish in habit, and crawled slowly, but
floated upon the surface foot uppermost.

This specimen died in the vessel of water, discolouring the fluid of a pink tinge/which,
however, did not appear to injure two Planariae in the same water.

Two figures of this species occur in Sir "VV. Elliot's 'Madras Nudibranchs '*. One of
these, with black markings on the mantle, is supposed to be the typical specimen, and one
of a more pure rose-colour the variety. I am disposed to think, however, that the
present description will be found typical, and that those diverging into markings are
varieties.

Family PHYLLIDIADvE, Lamarck (1809).
Genus Phyllidia, Cuvier (1708).

Phyllidia spectabilis, Coll., n. sp. (Plate X. figs. 19-23.)

Length 2 inches ; greatest breadth f inch. Body oval, tuberculated, covered with the
mantle. Tentacles two, short, placed near together, rather more than | inch from the
anterior extremity of the mantle, tapering, slightly curved, laminated, black, retractile
within a simple fixed sheath, which is situated on one of the tubercles of the mantle.

Upper surface covered with numerous irregularly-shaped tubercles, arranged in groups
of from one to ten or twelve, these clusters each perfectly distinct and similarly coloured.
The ground of the mantle is jet-black and smooth, forming a network which ramifies
among the groups of tubercles ; the bases of these groups polygonal, of a pale emerald-
green colour, the most elevated kuobs being whitish. A narrow, black, irregular edging
surrounds the dorsal surface, enclosing all the groups of tubercles, outside which is a
smoother and paler, irregular, and very slightly tuberculated margin. Under surface —
Foot greyish, oval ; a small bitentaculated head in front, the tentacles immovable. A
narrow cleft bisects the posterior half-inch of the foot. Mantle ample, surrounding the
head and foot on all sides. At the junction of the mantle with the foot and on the under
eck-e of the mantle is a close row of lamelliform branchiae, small in size, the series sur-
rounding the whole body, except the head.

This beautiful Phyllidia, in captivity, deposited a long, irregular, and narrow ribbon of
spawn, of a whitish colour, from an aperture in the side of the body. One specimen,
found under a block of coral between tide-marks on Pulo Barundum (or Marundum),
west coast of Borneo.

Doubts might be expressed whether this species may not come rxnder that named
Phyllidia ptistulosa, Cuv., P. verruculosa, Cuv., and Phillidiella pustulosa, Berghf. If

* Alder and Hancock, Trans. Zool. Soc. vol. v. p. 12G, pi. xxxi. figs. 1 and 2.

t See Cuvier, Ann. du Mus. (1804) vol. v. p. 268, pi. sviii. A. fig. 8 ; Mem. p. 3. fig. S ; also Bergh, Monog-
p. 511, and in Sernper's Eeisen im Arehip. d. Philipp. Band ii. Heft 10, p. 3^2.

NUDIBRANCHIATE MOLLUSCA FROM THE EASTEEN SEAS. 137

such should be the case, this Nudibranch has a wide distribution. As I cauuot, however,
myself regard theni as identical, I prefer to give that now described a separate specific
designation .

Genus Fryeria, Gray (1853).

Fryeria variabilis, Coll., n. sp. (Plate X. figs. 24-28.)

Length 2f inches ; breadth 1 J inch. Body convex, elongated, thickly tuberculated.
Tentacles two, situated f of an inch from the anterior margin of the body, arising from
simple sheaths, in which they are retractile ; the sheaths and tentacles black, the latter
being also finely laminated. Mai/ tie covering the head and sides of the foot, but leaving
tbe posterior point of the foot exposed during progression.

Colour. — The smooth ground-portion black, but supporting a number of irregularly-
scattered roundish tubercles, mostly single, a few double or treble ; the tubercles largest
and most loosely scattered over the central dorsal region, smaller and more closely
clustered along the sides. The colour of these tubercles was in one specimen pinkish, in
others of a pale emerald-green. In all respects, however, they are evidently the same
species. Foot smooth, black above where visible. Under surface. — Foot dark grey
beneath, blackish at the sides. Head small, with two small black tentacles concealed by
the mantle. Underside of the mantle blackish. A row of lamelliform branchige occu-
pying the junction of the foot and mantle beneath all round the body, excepting only
the head. Anus situated under the mantle on the right side, about f inch from the
anterior margin.

. Several specimens were found under blocks of coral on the reef of Palo Barundum,
west coast of Borneo. Animal very sluggish in its movements.

Family SCYLLJEIVJE, Alder & Hancock (1855).

Genus Scyll.ea, Linna3us (1758).

Scyll^a pelagicAj Linn. (Plate X. figs. 29-33.)

Length 1^ inch (average). Body smooth, opalescent, narrow, compressed, with two
broad tentacle-sheaths and two pairs of broad branchial lohes ; the posterior portion
raised to a crest, wedge-shaped, and notched like a cock's comb. Dorsal tentacles two,
lamellated, in broad and somewhat clavate compressed sheaths, straight in front and
finely serrated and crenatcd behind, with smooth and entire broad footstalks, the apex
somewhat flattened and depressed. Branchial lobes broad, flat, in two pairs, situated on
broad entire peduncles of nearly equal size ; externally smooth, internally having a
number of very delicate tree-like tufts irregularly scattered, of varying size, the large
below and the smaller above. Scad with a projecting crenated veil. Foot narrow, the
edges folding over inwardly, but when the animal crawls on glass a small portion of the
foot appears flattened out for the purpose.

Colour and general appearance. — General colour yellowish brown, darker along the
edges of the papillary prominences and the tentacles and their connecting ridges, inter-

138 DR. CUTHBERT COLLINGWOOD ON SOME NEW SPECIES OP

spersed here and there with minute white spots. Tufts on papillae [light brown. On
either side of the body is a row of opaque white projecting tubercles ; and between them
and tbe papillary prolongations are some minute turquoise spots, three or four in number.
The body generally is opalescent, with faint brown markings.

Spawn, a loose straw-coloured coil, entwining the leaves and berries of Sargassum
bacciferum, and imbedded in a mass of transparent jelly.

Portion of a branch of Sargasswm bearing a floating bladder, and with ($p) a coil of spawn of Scyllcea fdagica attached.

Nat. size.

Considerable numbers of this pelagic species were found upon the Sargassum floating
in lat. 25c X., long. 37° W., most pieces of the weed having one or more specimens. The
animals were in a constant movement of contraction and writhing. In the water they
swam freely, moving the head and tail from side to side alternately, so as nearly to touch
one another ; and when thus swimming were always, owing to the weight of the papillary
prolongations and tentacles, back downward, and bore a grotesque resemblance to a
four-legged animal with long ears (such as a Skye terrier). They would also attach
themselves to the surface of the vessel by a sort of sucker formed by a small cylindrical
portion of the foot (fig. 33, PI. X.). There is no figure of this species in Alder and
Hancock's work, nor do I know of any good figure. The present ones are from life, and
fig. 30, PL X. represents its peculiar falling aspect, as mentioned above.

Alder and Hancock (T. Z. S. v. p. 136) have described a Scyllcea marmorata and S. viridis
as new species, the former of which, save in size, differs little, if at all, from the Linnean
species S. pelagica ; but they admit that it is difficult to decide as to what constitute
specific differences in the genus. In Dr. Bergh's Monograph of Scyllcea* there is an
excellent resume of the subject, and a number of varieties of S. pelagica described and
referred to, as well as a good account of the anatomy of the genus given.

Family DENDRONOTLD^, Alder & Hancock (1855).

Genus Borxella, Gray (1817).

Bornella marmorata, Coll., n. sp. (Plate X. figs. 31-38 )

Body long and slender, with seven pairs of papillae. Head small, with a veil of five

* Semper's Reisen, 7. c, Band ii. Heft 8, p. 315 el sc</.

NTTDIBRANCHIATE MOLLUSCA FROM THE EASTEEN SEAS. 139

or six radiating papillae on either side, deeply cleft in the middle. Dorsal tentacles
situated upon thickish footstalks, from which spring one large hranch (the uppermost)
and four shorter branches, forming altogether a radiating sheath, in the midst of which
is a slender laminated tentacle. Marginal processes in seven pairs, the anterior longest
and most compressed, the posterior pair small and rudimentary. Each pair of papilla?
varies in size and development : the first pair consists of three elongated bodies of the
size of the largest branch in the radiating sheath of the tentacles ; the second pair has
two such ; the third and fourth also, but smaller ; the last pair has but one. On the
inner side of each of these bundles is a finely ramified dendriform branchia, not so large
as the processes, the processes and branchia all arising from a common stem. Foot long,
and pointed behind. Anus situated on the right side, between the dorsal tentacles and
first pair of papillae.

Colour and general appearance. — Body beautifully marbled all over with streaks of
vermilion, strongest on the back, and more delicate and faint on the sides of the foot.
Papillae tipped with vermilion, and also the radiating projections of the veil, but neither
are marbled like the body. Dorsal tentacle-branches also tipped with vermilion, and the
largest branch marbled.

Three specimens of this beautiful species rewarded a brief search among the rocks near
the landing-place at Aden in March. They were under stones, not far from high-water
mark. The animals were extremely active, often swimming with a lateral vermicular
movement, at which time the body was laterally compressed. They were seldom quite
at rest.

I here specifically distinguish this form, although previously * I inclined to regard it
as the B. digitata, Adams t ; but a more careful examination since discloses points of
difference which had previously escaped my notice. B. digitata was first found at Sunda
on floating fuci, but it has since been got on the Madras coast by Sir Walter Elliot J, and
in the Philippines by Professor Semper §. Bergh § enumerates six species belonging to
the genus, and he adds very considerably to the account of the anatomy furnished by
our countrymen Messrs. Alder and Hancock ||.

DESCRIPTION OF THE PLATES.

The line beside certain of the figures denotes the natural size of the animal. The separate figures of
tentacles, branchial tufts, and other parts are all considerably enlarged.

Plate IX.

Figs. 1-5. Doris pecten, Coll., n. sp. : 1, general view of animal, its dorsum; 2, under surface of same ;
3, a dorsal tentacle (rhinophore) ; 4, a branchial tuft ; 5, outline of the animal at rest.

* " Observations on the Distribution of some Species of Nudibranchiate Mollusca in the China Sea," Ann. & Mag.
Nat. Hist. 1868, ser. 4, vol. i. p. 91 ; and ' A Naturalist's Rambles in the China Seas' (London, 1868), p. 9.

t Voyage of the ' Samarang," p. 07, pi. six. fig. 1. X Trans. Zool. Soc. vol. v. p. 131.

§ Iteisen im Archip. d. Thilipp. Band ii. Heft 7, p. 301. || Loc. cit. T. Z. S. swpra.

SECOND SEKTES. — ZOOLOGY, VOL, II. 19

140 ON SOME NEW SPECIES OF NUDIBRANCHIATE MOLLUSCA.

Pigs. 6-8. Doris crescentica, Coll., n. sp. : 6, general view of animal, its dorsum (nat. size) ; 7, a dorsal

tentacle and its sheath, anterior view ; 8, a branchial tuft (or branch). Sketch taken from

specimen obtained at the island of Barundum, west coast of Borneo.
Figs. 9-14. Chromodoris iris, Coll., n. sp. : 9, lateral view of animal in progression, the branchiae partly

closed ; 10, another view of the same, with the branchiae fully expanded ; 11, a side view of the

animal contracted and at rest ; 12, a lamina of the branchiae ; 13, a dorsal tentacle ; 14, the

spawn.
Figs. 15-17. Chromodoris Bullockii, Coll., n. sp. : 15, the animal in progression; 16, the underside of

the head; 17, plan in outline of the brancbial laminae.
Figs. 18-22. Chromodoris aureo-purpurea, Coll., n. sp. : 18, dorsal surface of the animal in progression;

19, outline from above of the animal at rest ; 20, diagram of the mode of arrangement of the

branchiae ; 21, a branchial lamina; 22, a dorsal tentacle.
Figs. 23-26. Chromodoris tumuli/era, Coll., n. sp. : 23, dorsal surface of the animal in progression; 24,

a tubercle of the dorsum seen in profile; 25, a branchial lamina; 26, a dorsal tentacle.
Figs. 27-29. Chromodoris tenuis, Coll., n. sp. ; 27, upper surface of the animal in progression; 28, the

under surface of the animal as seen swimming with foot ; 29, anterior view of a dorsal tentacle.
Figs. 30-33. Chromodoris funerea, Coll., n. sp. : 30, animal in progression as seen from above; 31, a

dorsal tentacle; 32, a branchial tuft ; 33, an outline plan of the branchiae.
Figs. 34-37. Chromodoris Alderi, Coll., n. sp. : 31, dorsal surface of the animal in progression ; 35, larnime

of the branchiae ; 36, a dorsal tentacle ; 37, the suctorial mouth.

Plate X.

Figs. 1-5. Albania formosa, Coll., nov. gen. et sp. : 1, animal in movement as seen swimming, its upper

surface ; 2, the veil ; 3, a tuft of the branehue ; 4, a dorsal tentacle ; 5, an outline of the

under surface.
Figs. 6-11. Triopa Principis-JVallue , Coll., n. sp. : 6, animal in progression, upper surface, the branchiae

expanded ; 7, side view of the animal at rest ; 8, a papilla of the dorsum ; 9, a papilla of the

head ; 10, a dorsal tentacle ; 11, outline of a branchial twig.
Figs. 12-14. Trevelyana felis , Coll., n. sp., immature : 12, animal in progression ; 13, animal as contracted

at rest; 14, natural size.
Figs. 15-17. Doridopsis arborescens, Coll., n. sp. : 15, the animal in progression ; 16, outline of the under

surface of the body ; 17, a branchial tuft.
Fig. 18. Doridopsis rubra, Kel. : diagram of the branchiae.
Figs. 19-23. Phyllidia spectabilis, Coll., n. sp. : 19, upper surface of the animal ; 20, outline of the under

surface of the same; 21, a dorsal tentacle and sheath; 2.2, diagram portion of branchiie ; 23,

the spawn.
Figs. 24-28. Freyeria variabilis, Coll., n. sp. : 24, animal, its upper surface; 25, diagram of the under

and lateral surfaces, showing position of branchiae ; 26, an under view of the head ; 27, a dorsal

tentacle ; 28, sketch portion of the branchiae.
Figs. 29-33. Scylhea pelayica, Linn. : 29, dorsal surface of the animal; 30, position which the animal

assumes in falling through the water; 31, a dorsal tentacle sheath; 32, branchial tuft covering

the inner surface of the body ; 33, the foot from below.
Figs. 34-38. Bornella marmorata, Coll., n. sp. : 34, outline sketch of the head; 35, a dorsal teutacle;

36, the veil ; 37, a branchial lobe ; 38, markings on the mantle.

[ 141 1

111, On the Anatomy of Ants. By Sir John Lubbock, Bart, MB F.B S F.L S
B.CL LL.B Vice- Chancellor of the University of London, President of 'the
Entomological Society. J

(Plates XI. & XII.)

Bead February 6th, 1879.

Introductory Bemarks.

IN conjunction with the observations on the habits of Ants, which the Society has done

me the honour of publishing from time to time in the Journal, I have also been studying

their anatomy especially with reference to the muscular system. Of the anatomy of the

Society '1877* **** ^ ^^^ ™ ^ ' Tnm8action8 of the Microscopical

The present paper is devoted to the thorax, with special reference to Basins favus.
Though it is founded on numerous dissections, and on more than 1000 sections beautifully
prepared for me by Mr. Newton, of the School of Mines, and Mr. Robertson, of Oxford it
is still very imperfect ; and I am only induced to bring it before the Society in its present
incomplete state because, while I hope it will be found to add somewhat to our knowledge
I see little prospect of being able to work out the subject as thoroughly as I could wish '
As a general rule, the thorax of insects is considered to consist of three more or less
well-marked segments, usually known by the names suggested by Nitzsch-prothorax
mesothorax, and metathorax. r '

Dr. Ratzeburg, however, published in 1832 a memoir (< Ueber Entwickelung der fuss-
losen Hymenopteren-Larven, mit besondercr Riicksicht auf die Gattung Formica ') in
which he maintained that the fifth segment of the larva forms, not the so-called « scale >'
or firstabdommal segment, but the hinder part of the thorax. This view has also been
maintained by Audouin and Latreille; while, on the contrary, others, as, for instance,
Kirby and Spence and MacLeay, consider the thorax of these insects to be composed
ot three segments, as usual.

Burmeister, indeed, roundly observes (< Manual of Entomology,' Shuckard's transl.
p. b5J that Audouin s assertion is unfounded.

Lepelletier cle St.-Fargeau, in his < Histoire Naturelie des Hvmenopteres ' (183G)
expresses the same opinion. « II me parait," he says, "plus simple parler comme voient

7? y6T • n, m n alS°' ^ MS eXCdlent ' Introd^tio» to the Modern Classification

of Insects (1810, p. 227), adopts the same view. It may, he admits, « be asserted that,
as the body of the imago possesses two or three segments fewer than exist in the body of
larva, we may suppose that the loss of one of these segments takes place at least in this
manner, and in this part of the body. This, however, can only be done by admitting that
the head and three thoracic segments of the imago are composed of five larvta-se.-
ments instead of four, an admission negatived by all analogy with pedate larva,."

* See Quekett Lecture, Monthly Microscopical Journal (1877). vol. xviii. p. 121.

19

*

142 Sffi JOHN LUBBOCK ON THE ANATOMY OF ANTS.

Newport, in his article " Insecta," in the ' Cyclopaedia of Anatomy and Physiology,'
p. 920, says : — " at first we were inclined to Audouin's opinion, more especially on
account of what we shall presently find in Lepidoptera, in which the fifth segment, in its
atrophied condition, is as much connected with the thorax as with the ahdomen. On
further examination, however, we are satisfied that that portion of the metathorax
which is posterior to the incisure belongs to the third segment of the thorax."

Fenger, in his " Allgemeine Orismologie der Ameisen " (Arch, fur Naturg. 1862, p. 315),
treats the thorax as consisting of three segments, and does not even hint at any dif-
ference of opinion on the subject. Mayr also, in his excellent ' Die europaischen For-
miciden,' p. 4, and Schenck, in his " Bcschreibung nassauischer Ameisen- Arten " (Jahrb.
des Ver. fur Naturkunde im Herz. Nassau, 1852), adopt the same view. Lastly, Forel
(' Fourmis de la Suisse,' p. 5) says that the thorax " se divise en trois segments, comme
chez tous les insectes : prothorax, mesothorax et metathorax."

It would perhaps hardly be appropriate to refer to more general or condensed works
in which the thorax is stated to consist of three segments, as, for instance, by Owen
(' Lectures on Invertebrate Animals,' p. 193), Claus (' Grundziige der Anatomie,' p. 557),
Bolleston (' Forms of Animal Life,' p. cix), &c. ; for these eminent authors, thougb expres-
sing no qualification, perhaps only meant to describe a general, and not necessarily an
invariable, rule.

Huxley, in the ' Introduction to the Classification of Animals,' p. 58, observes, with his
usual care and accuracy, that " three, or perhaps, in some cases, more, somites unite, and
become specially modified to form the thorax."

Notwithstanding the high authorities who have adopted the opposite opinion, and
although the first appearance of the tborax seems to support their view, for my own part
I cannot but think that Batzeburg's opinion was correct. Packard (' Guide to the Study
of Insects,' p. 66) has given figures of the metamorphoses of Bombus, from which it seems
clear that the fifth segment of the larva forms the posterior portion of the thorax of the
perfect insect. Lacaze-Duthiers (Ann. des Sc. Nat. 1853, p. 231), Palmen (Zur Morph. des
Tracheensystems), and Beinhard (Berl. ent. Zcits. 1865) also advocate the same view.

The position of the spiracles affords also strong evidence in support of the same opinion.
It is generally stated in works on the anatomy of insects that there are on the thorax
two pairs of spiracles, the first between the pro- and mesothorax, the second between the
meso- and metathorax.

According, indeed, to Burmeister (' Handbook of Entomology,' p. 164), this is also the
case with the Hymenoptera, which " all possess four (spiracles) in the thorax, two of which
are upon the limits of the prothorax, between it and the mesothorax, and the other two
lie between the meso- and metathorax. In the Hymenoptera, in which the thorax con-
sists of a hard, horny case, and the segments are closely united together, the posterior
pair of spiracles lie upon the metathorax itself, whereby they distinguish themselves from
all the other orders." In fact, however, as may be seen from the accompanying
figures (PI. XL figs. 2, 4, 5), the thorax of Ants possesses, not two, but three, pairs of
spiracles.

Tbe two first pairs are situated between the pro- and mesothorax and the meso- and

SIR JOHN LUBBOCK ON THE ANATOMY OF ANTS. 143

metathorax, as usual, and evidently correspond with the two pairs of thoracic spiracles of
other insects. The third pair is situated at the side of the so-called metathorax ; but in
no case whatever do we find among insects two pairs of spiracles on one segment. Such
an arrangement would be contrary to the whole plan of organization of the Arthropoda.
It is obvious, therefore, that the third pair of spiracles corresponds to that which in other
insects lies between the thorax and the first abdominal segment. Burmeister, as we
have seen, remarks that certain Hymenoptera " distinguish themselves from all other
orders " in having a pair of spiracles " on the metathorax itself;" but he supposes that
these correspond to the spiracles which are ordinarily situated between the meso- and
metathorax, overlooking the fact that these spiracles also exist as usual. It seems clear,
therefore, that the portion of the body posterior to the third pair of spiracles really cor-
responds to the first abdominal segment in ordinary insects.

Nor are the respiratory organs alone in pointing to this conclusion. The internal
chitinous appendages clearly divide the thorax into four portions ; and I think it may
be said that the thorax contains four ganglia, though the last (PI. XL fig. 2, G4, PL XII.
fig. 2) is certainly not large.

The Prothorax.

The upper part of the prothorax, or pronotum, is formed in Lasius jlacus by a single
arched chitinous plate (PL XL figs. 1, 1, & 6, B), which slopes downwards from its posterior
border towards the head, where it forms a sort of keel (PL XL figs. 1 & 2). Seen exter-
nally and from the side, its lower border appears to join the upper edges of the propectus ;
but a transverse section (PL XII. fig. 4) shows that this is not so, but that the propectus
is continued for some distance beyond the lower margin of the pronotum, and is then
connected with it by a membrane which passes from the upper margin of the propectus to
the lower one of the pronotum: The propectus tapers in front (PL XL fig. 5), terminating
on each side in two teeth, which lock into two corresponding teeth (PL XL figs. 6 & 7,
and PL XII. fig. 1, X) or processes at the back of the head. The propectus is divided
into two plates (Plate XL fig. 2, C & T), one anterior and one posterior, which, moreover,
are divided into lateral regions by a central ridge. The anterior plate of the propectus
has in front a deep bay or depression, at the two horns of which are the above-mentioned
teeth or processes. Each region of the anterior division of the propectus has therefore
roughly the form of a triangle with arched sides. The posterior division of the pro-
pectus is elliptic in form, and not so large as the anterior division, to which it is firmly
attached.

The propectus is therefore attached to all the surrounding chitinous plates by flexible,
though tough, membranes. It hangs, indeed, something like the under body of a carriage;
xnd from the fact that the anterior horns of the prothorax interlock with the posterior
arocesses of the head, if the propectus is turned round it carries the head with it. On
;he other hand, if the head be retracted, the posterior processes of the head, from their
position with reference to the anterior horns of the prothorax, prevent the head of the
nsect from being turned round against its will.

144 SIE JOHN LUBBOCK ON THE ANATOMY OF ANTS.

The posterior surface of the propectus is connected with the anterior edge of the meso-
pectus hy a tough, hut flexible, membrane.

I have found it difficult to understand the descriptions given of the interior skeleton
of the thorax hy previous writers, nor do their figures give much assistance. In the
normal insect-thorax there appear to he seven principal processes — four springing from
the back, and called by Kirby and Spence the phragma, prophragma, mesophragrna, and
metaphragma ; and three from the sternum, named by the same authors profurca,
mesofurca, and postfurca. In the worker Ants the four superior processes are not
developed, but the furca, mesofurca, and postfurca are very important ; they give
attachment to various muscles, and serve also to protect the nervous system. Kirby and
Spence, however, dismiss them very summarily, and, as regards the processes of the endo-
sternum, state that they "are not sufficiently remarkable to require particular notice " *.
Burmeisterf says that in the prothorax (of the Hymenoptera) "there are two strong
pointed processes, each of which has a double root. The exterior one comes from the
margin of the prosternum, and the interior one from the central ridge of the same part.
Between these roots the muscles of the coxae pass, and between the processes themselves
run the pharynx and the nervous cord ; and it is to these processes that the connecting
muscles of the pronotum and prosternum are attached. In the mesothorax we first rind
the prophragma, a small, not very high, horny partition, which descends from the anterior
margin of the mesonotum ; and we next find a delicate ridge, which encompasses the
whole distinctly separated mesonotum. The mesosternum and scapuke are closely joined
in a half-ring, and from the central carina of this ring springs a broad strong ledge, which
at its upper margin is furnished on each side with a strong process ; they form with the
ledge a rectangular cross, and serve as points of insertion for the muscles of the coxa- of
the middle legs, lying on each side contiguously to the central ridge." As regards the
metafurca, all he tells us is, " between the metanotum and metaphragma the two large side
pieces and their auxiliaries lie, separated from each other by furrows, from which, inter-
nally, strong ridges spring, and to which the muscles of the posterior legs are attached."

Graber, in his admirable ' Die Insekten,' truly observes that the endoskeleton has been
almost entirely neglected by recent entomological writers. I trust, however, that the fol-
lowing description and the accompanying figures may give some idea of the endoskeleton
as it exists in the workers of Lasiusjlaciis.

The hinder plate of the propectus turns upwards at approximately a right angle, and
is produced into the antefurca (PI. XI. figs. 1, 2, 5, & 6; PI. XII. fig. 8), a chitinous pro-
cess which extends more than halfway up the dorsum, leaving, however, a central orifice
(PL XII. fig. 4) through which the nervous chords penetrate, while the oesophagus and
the heart pass between the upper edge of the antefurca and the dorsum.

As seen from behind (PI. XII. fig. 4) it has the form of a cross with four arms. In the
middle of the centre piece is an oval orifice, the wider end below, through which the
nervous system passes. The centre of the upper part sends out a process both anteriorly
and posteriorly, as shown in PL XL fig. 2 ; in fact it forms a sort of case for the protection
of the ganglia.

• ' Introduction to Entomology,' vol. iii. p. 587. t Op. cit.

SIE JOHN LUBBOCK ON THE ANATOMY OF ANTS. 145

The medifurca (PI. XI. fig. 2 ; PL XII. fig. 5, Med) rises from the medipectus. It is
much more elongated and slender than the antcfurca, and has the form of a Y, the upper
arms of which, however, are connected by a cross bar, thus leaving a triangular orifice
with rounded angles, through which runs the nervous chord. To a process of the cross
har is attached the muscle which elevates the prothorax.

The postfurca (PI. XL fig. 2 and PL XII. fig. 6) also has somewhat the form of
a Y. The stem, however, is much shorter, the branches are curved, and the cross bar
is absent. The postfurca arches forwards, so that the upper part of the arms approach
those of the medifurca, with which they are connected by tendinous fibres. Between
the medifurca and the postfurca lies the third thoracic ganglia.

Muscles of the Mead.

There are two elevators of the head on each side (a & a1). The first (PL XL figs.
1, 2, & 5, a) is a thin muscle, which rises from the back near the middle line, at the
junction of the pro- and mesothorax, and, passing forwards, is inserted at the upper margin
of the occipital foramen, where the posterior margin of the head joins the intersegmental
membrane. The second is more powerful. It (PI. XL figs. 1, 2, a1) rises from the anterior
surface of the upper part of the antefurca, and, passing forwards and slightly upwards, is
inserted close to the preceding. The heads of attachment of this muscle reach almost
across the segment.

The first depressor of the head (PI. XL figs. 1, 2, b), like the second elevator, is attached
to the anterior face of the antefurca, but at a lower level, and, passing over the prothoracic
ganglion, is attached to the inferior margin of the occipital foramen.

The second depressor of the head (PL XL fig. 1, PL XII. fig. 1, bx) is attached to the
central and hinder part of the propectus, and, passing directly forwards, is also attached
co the lower edge of the occipital foramen.

The rotators of the head are five in number on each side. The first (PL XL figs. 1, 2, &

t,c) rises from the middle of the lateral wall of the pronotum, and, passing downwards

ind inwards, is attached to the anterior toothed process of the propectus. The second

'otator passes from the middle of the lateral wall of the propectus (PL XL figs. 4, 6, &7, cl),

md is attached to the outer anterior toothed process of the prosternum. The third

■otator (PL XL figs. 4, 6, 7, c'2) lies rather nearer the middle of the segment. In front it

s attached to the inner toothed process, and posteriorly to the lateral and posterior

vail of the propectus, a little behind the preceding. The fourth rotator (PL XL

igs. 1, 2, 6, d) commences at the anterior process of the propectus, close to the preceding,

md, passing backwards and slightly inwards, is attached to the anterior central process

>f the antefurca. The fifth rotator (PL XL figs. 1, 2, & 6, dl), rises with the preceding,

mt passes diagonally across the segment to be attached to the lateral edge of the antefurca.

Although the muscles of the head of Coleoptera, as described by Straus-Durckheim in

Melolontha, and as given generally by Burmeister in his ' Handbook of Entomology,' are

aore complex than those which are found in Ants, yet neither of these authors describe

ny muscle exactly comparable to the following.

This muscle (PL XL figs. 1, 5, e) differs from the preceding in that, while they taper

146 SIE JOHN LUBBOCK ON THE ANATOMY OF ANTS.

as they pass forward, it, on the contrary, rises from the anterior surface of the pronotum
by several, somewhat diverging heads, and, passing backwards and slightly downwards,
is attached to the upper part of the antefurca. It would therefore seem to draw the
propectus, and consequently to push the head, forwards. It is obvious that if the head
is projected forwards, and the propectus then retracted, so that the head could move
freely towards each side, it would be easily turned by the rotators above described. On
the contrary, if it be retracted, or if the propectus be thrown forward, so that the
posterior process of the head interlocks with the anterior processes of the propectus,
the head would be so situated as to retain its position even against a considerable force.

The next muscles to be mentioned are the elevators of the antepectus; these are
two in number. The first (PI. XL figs. 1, 2, 4, 5, and PL XII. fig. 4,/) rises from
near the middle of the pronotum, and, passing downwards, is attached to the anterior
process of the antepectus. The second is weaker; it is attached to the side of the
pronotum, and, passing downwards (PL XL fig. 5,/1) and slightly inwards, is also
attached to the anterior process of the antepectus, close to the preceding. The attach-
ment of the first large rotator of the head (e) lies between those of these two muscles,
as may be seen in PL XL fig. 5, where f and fl represent the heads of these two
muscles, which, when they contract together, would tend to elevate the antepectus.

The depressor of the antepectus is smaller. It commences (PL XL figs. 5, 6, 7, g) at
the lower edge of the pronotum, and, passing upwards, is attached to the upper edge of
the antepectus, which therefore, on contracting, it draws downwards.

Front Legs and their Muscles.

The legs consist of the following segments : — 1, coxa ; 2, trochanter ; 3, femur ; 4, tibia ;
and 5, tarsus, this latter being composed of five segments.

The description given by Straus-Durckheim of the muscles by which the legs are
moved has been adopted by most subsequent writers. According to him, the anterior
le°'s have five muscles, four flexors and one extensor. The first flexor rises from the
superior lateral and anterior surface of the prothorax, and is attached to the posterior
border of the coxa. The second and third flexors rise from the superior and posterior
surface of the prothorax, and are attached to the coxa just outside the preceding. The
fourth flexor rises from the external portion of the posterior surface of the " rotule,"
and is attached to the posterior edge of the coxa. Lastly, the extensor rises from the
pronotum, near the first flexor, and acts immediately in opposition to the preceding.

The number of muscles in the Ant appears to be greater than in Melolontha, audi
the disposition is in many respects dissimilar.

The first muscle of the leg (PL XL figs. 4, 7, and PL XII. figs. 1, 4, h) rises from the I
anterior lateral wall of the prothorax, and, passing downwards and backwards, is
attached to the upper anterior angle of the condyle of the coxa, which, therefore, it
would tend to draw forwards and inwards.

The second (*, PL XL figs. 1, 2, 4 ; PL XII. figs. 1, 3, & 4) lies transversely in the lower
and posterior portion of the antepectus. In PL XL figs. 1 & 2 it is seen in section. In
PL XL fis?. 4 it is severed close to its attachment. It rises from the central ridge

o

SIK JOHN LUBBOCK ON THE ANATOMY OF ANTS. 147

of the antepectus, and, passing trausversely across the segment, is attached to the
posterior and outer edge of the leg, at the summit of the projecting head or condyle.
It would tend to extend the leg laterally.

The third (i1, PL XII. fig. 4) is attached to the antefurca, and, passing downwards
and outwards, is attached close to the preceding.

The fourth and fifth muscles of the fore legs are of a different character, penetrating
into the coxa. The fourth rises from the upper edge of the antepectus in front of the
antefurca (PL XL fig. 4, and PL XII. fig. 4, k), and passes downwards into the coxa.

The fifth rises partly from the hinder wall of the antefurca, partly from its posterior
spur (PL XI. figs. 1, 2, 4, & 7, /), and, like the preceding, passes down into the coxa.
The upper part of the muscle is joined by some fihres, which pass round the posterior
process of the antefurca and are attached to the pronotum.

The seventh is attached to the outer and posterior edge of the coxa, and, passing
backwards and inwards, is attached to the anterior surface of the medifurea. It is not,
however, well shown in any of my sections.

In addition to these muscles, the coxa contains two others, one of which rises from the
upper and outer wall and passes downwards and inwards, while the other, rising from the
upper and inner wall, passes downwards and outwards.

The small trochanter (PL XII. fig. 1, tr.), in addition to the above-mentioned fibres of
the flexor of the femur, contains only a short single muscle, which at its lower end is
ittached to the thigh.

The femur (PL XII. fig. 1, fm) contains two muscles. The extensor is attached to
she upper surface of the segment, the fibres being attached to one side of a long tendon,
svhich at its lower end is attached to a chitinous piece at the upperside of the head of
;he tibia. The flexor is situated rather on the lower side of the segment ; but the fibres
liverge from both sides of the tendon, and some of them cross those of the extensor
nuscles. Some of the central fibres pass into the trochanter, and are attached to its
nner margin. The lower end of the tendon of the flexor is attached to a chitinous
irocess.

The tibia presents some very remarkable points, with reference to which I may perhaps
)e permitted to quote a passage from a paper of mine published in the ' Microscopical
Journal,' 1877.

Remarks on the Tibial Organ.

In the year 1844 Von Sicbold* described a remarkable organ which he had discovered
n the tibiae of the front legs of Gryllus, and which he considered to serve for the purpose
|)f hearing. These organs have been also studied by Burmeister, Brunner, Hensen,
^eydig, and others, and have recently been the subject of a monograph by Dr. V. Graberf,
vho commences his memoir by observing that they are organs of an entirely unique
haracter, and that nothing corresponding to them occurs in any other insects or, indeed,
n any other Arthropods.

* " Ueber das Stinim- und Gehor-Organ der Orthopteren," Wicgmann's Arch. f. Natur. 1S44.
t Die tympanalen Sinnes-Apparate dcr Orthopteren, von Dr. Titus Graber, 1875.
SECOND SERIES. — ZOOLOGY, VOL. II. 20

148 SIR JOHN LUBBOCK ON THE ANATOMY OF ANTS.

I have therefore been very much interested by discovering in Ants a structure which
seems in some remarkable points to resemble that of the Orthoptera. As will be seen
from a glance at Dr. Graber's memoir, and the plates which accompany it, the large
trachea of the leg is considerably swollen in the tibia, and sends off, shortly after entering
the tibia, a branch, which, after running for some time parallel to the principal trunk,
joins it again. See, for instance, in his Monograph, pi. ii. fig. 43, pi. vi. fig. 09, pi. vii.
fig. 77, &c. Now I have observed that in many other insects the tracheae of the tibia
are dilated, sometimes with a recurrent branch. The same is the case even in some mites.

I will, however, reserve what I have to say on this subject, with reference to other
insects, for another occasion, and will at present confine myself to the Ants. If we
examine the tibia, say of Lasius Jlaviis, we shall see that the trachea presents a remark-
able arrangement, which at once reminds us of that which occurs in Gryllus and other
Orthoptera. In the femur it has a diameter of about 3<fo5 °f an mcn > as soon, however,
as it enters the tibia it swells to a diameter of about -5-^ of an inch, then contracts again
to -g^o, and then again, at the apical extremity of the tibia, once more expands to -^>q.
Moreover, as in Gryllus, so also in Formica, a small branch rises from the upper sac,
runs almost straight down to the tibia, and falls again into the main trachea just above
the lower sac. The remarkable sacs at the two extremities of the trachea in the tibia
may also be well seen in other transparent species, such, for instance, as Myrmica rugi-
nodis and JPheidole megacephala.

At the place wdiere the upper tracheal sac contracts there is, moreover, a conical
striated organ (x), which is situated at the back of the leg, just at the apical end of the
upper tracheal sac. The broad base lies against the external wall of the leg, and the
fibres converge inwards. In some cases I thought I could perceive indications of bright
rods, but I was never able to make them out very clearly. This also reminds us of a
curious structure which is found in the tibia of Locustidaa, between the trachea, the
nerve, and the outer wall, and which is well shown in some of Dr. Graber's figures.

Lb'

Other Organs of the Prothorax.

The anterior pair of spiracles, as already mentioned, lie (PL XL figs. 4, 5, Spl) between
the pro- and mesothorax. The tracheal tube immediately behind the spiracle is provided
with a short muscle, as already described in other insects by MM. Landois and Thelen*.
The ganglion (PI. XL figs. 2, 6, & 7, Gl) is of considerable size, and is connected ante-
riorly with that of the head, and posteriorly with that of the mesothorax, by a double
commissure. In the latter case the commissures pass through an orifice in the antefurca,
which thus not only serves as a support to the muscles, but also as a protection to the
nervous system.

The oesophagus passes straight through the prothorax, and, indeed, does not enlarge
into the crop until it reaches the enlarged part of the abdomen. In the upper part of the
prothorax lie the large thoracic salivary glands (PI. XL fig. 2, gl).

A considerable part of the upper and anterior portion of the prothorax is occupied by
the thoracic salivary glands, which I have already described in the ' Microscopical Journal.

* Zeitsehr. f. wiss. Zool. 1867, p. 1S7.

SIR JOHN LUBBOCK ON THE ANATOMY OF ANTS. 149

They consist of a number of branched and twisted tubules which gradually unite in a
single duct. This duct then swells into a capacious receptacle, after which it contracts
again, and after joining the corresponding duct from the other side, passes through the
neck into the head, and then, after a meandering course, opens at the upperside of the
under lip. The duct consists of an epithelial layer of cells, within which is a structureless
membrane, strengthened, as is so often the case with the ducts of glands, by chitinous
ridges, which give it very much the appearance of a trachea. Fig. 3, PI. XI., represents
a glandular organ situated in the lower part of the thorax of Myrmica ruginodis imme-
diately above the base of the anterior leg.

Mesotliorax and Middle Legs.

The mesotliorax is much more closely connected with the metathorax than with the
prothorax (PL XI. fig. 2). Like the prothorax it consists of an upper and lower more or
less arched plate. The upper plate or mesonotum (PI. XL figs. 2, 5, 6, lies) is oblong,
somewhat emarginate behind, the spiracles (PI. XL figs. 2, 4, Sj)2) being situated at the
posterior angles. In front the mesonotum projects some way over the sides of the pro-
thorax ; and as the middle legs are attached quite at the posterior end of the metapectus,
they, as well as the posterior legs, lie under the metanotum, and seem at first sight as if
they belonged to the hinder division of the thorax.

The depressor of the prothorax (PI. XL figs. 2, 4, 5, in) arises from the junction of the
meso- and metathorax, beneath the spiracle, and passing down and forwards is attached
to the lower posterior edge of the prothorax, which therefore it would tend to draw
downwards.

On the other hand, the elevator of the prothorax (PL XL figs. 2, 5, n) rises from the
upper part of the antefurca, and passes backwards and downwards to a spur of the
medifurca just above the mesothoracic ganglion.

The second pair of legs has, according to Straus-Durckheim's description of Melolontha,
three flexors and two extensors. The arrangement, however, is very different from that
in the Ant.

Graber, in his excellent work, refers specially to four muscles ; the first (irfm, in his
fig. Gl) rises from the central ridge of the sternum, and, passing directly outwards, is
attached to the inner edge of the coxa, Avhich therefore it would draw inwards and
downwards. The next two (slim 1 and 2, in his fig. 61) rise one behind the other from
the side of the thorax, and would, on the contrary, draw the leg outwards and upwards.
The fourth also lies behind the other two, but would specially draw the leg upwards.

As regards the Ant, the principal muscles which move the middle legs are shown
in PL XL fig. 4, PL XII. fig. 2.

The first muscle (o, PL XL figs. 4, 5, and PL XII. figs. 2 & 5) rises partly from the
upper lateral wall of the mesonotum immediately under the spiracle, partly from the
medifurca, and passing downwards contracts into a tendon which is continued into the
leg. It would tend to raise the leg.

The second rises from the anterior edge of the medipectus (PL XII. fig. 2, p), and
passing straight back is attached to the anterior edge of the coxa.

20*

150 SIR J01IN LUBBOCK ON THE ANATOMY OF ANTS.

In opposition to this the third muscle rises from the anterior portion of the central
ridge of the medipectus (PI. XII. fig. 2, q), and passing outwards and backwards is attached
to the inner posterior edge of the coxa.

The fourth rises from the posterior portion of the central ridge of the medipectus, and
passing outwards is attached (PI. XII. fig. 2, r) to the inner edge of the coxa.

The fifth rises partly from the anterior wall of the medipectus (PL XII. fig. 2, s), partly
from its median ridge, under q, and is attached to the outer anterior edge of the coxa.

The last (PL XII. figs. 2, 5, t) rises from the medifurca, and passing downwards and
forwards is attached to the outer edge of the coxa.

Posterior Portion of Thorax.

The first elevator of the abdomen (it, PL XI. figs. 2 & 5) rises from the metanotum, on
each side of and not far from the central line, and, running parallel to the same muscle
on the other side, is attached to the upper anterior edge of the so-called knot.

The second elevator of the abdomen [u\ PI. XI. fig. 2, PI. XII. fig. 2) rises from the
postpectus, and passing upwards, outwards, and backwards is attached to the upper
lateral anterior edge of the abdomen. It would draw the abdomen upwards and at the
same time sideways.

The depressor of the abdomen rises partly from the metanotum behind the first
elevator, and partly (v, PI. XL fig. 2, PL XII. fig. 2) from the upper part of the postfurca,
and passing backwards and downwards is attached to the lower anterior edge of the
abdomen.

The rotator of the abdomen rises from the metanotum just behind the first elevator
(tv, PL XL figs. 2, 5, and PL XII. fig. 2), and passing backwards, downwards, and out-
wards is attached to the lateral edge of the first abdominal segment.

I now pass to the muscles of the posterior leg.

The first muscle of the leg rises partly from the lateral wall of the metanotum
(x, PL XL fig. 5, PL XII. fig. 2) and partly from the upper part of the postfurca, and
passes downwards and backwards into the coxa.

The second muscle also rises from the postfurca below the preceding (PL XII. fig. 2, xx),
and passing downwards and backwards is attached to the upper posterior margin of the
coxa. It terminates above in a strong chitinous tendon, which is connected with the
postfurca by a number of tendinous filaments.

The third muscle rises from the lateral wall (PL XII. fig. 2, y) of the mesothorax,
partly from that of the metathorax, and passing backwards is attached to the outer edge
of the leg.

The fourth muscle is attached to the anterior edge of the postpectus (PL XII. fig 2, yl),
and passing backwards and outwards is attached to the external margin of the leg close
to the preceding.

The fifth is also attached to the anterior edge of the postpectus, but, passing directly
backwards (PL XII. fig. 2, z) above the preceding is attached to the exterior and anterior
margin of the leg.

The sixth muscle is attached to the anterior edge of the metathorax, and passes

SIE JOHN LUBBOCK ON THE ANATOMY OF ANTS. 151

directly backwards (PI. XII. fig. 2, z1) and over the preceding to the internal and posterior
edge of the posterior leg.

Although the workers of Ants do not possess wings, Dcwitz has shown * that the larva;
possess " imaginal disks," like those from which the wings of the males and females are
developed, but smaller. These embryonic wings reach no more advanced stage than that
which they have already acquired in the full-grown larva, and in the imago no trace of
the front wings appears to be discernible, while it is curious that the hinder wings, though
they are smaller in the males and females, are in some cases still indicated by a minute
protuberance.

The presence of wings necessarily entails may other differences, and consequently

The thorax of the male and female Ants is very unlike thai of the workers — not, indeed
in the arrangement of the muscles already described, bat by the changes and additions
contingent upon the presence of wings. The females, as is well known in most cases,
strip off their wings soon after the marriage-flight. In Anergates atratulus the males are
wingless, and, according to Schenckf, the queens in some cases do not acquire win^s. The
great muscles of flight are, as might be expected, very large in the winged Ants ; on the
other hand, they are few in number, more simple, as it would appear, tban those of most
other insects. There are, indeed, several small muscles attached to the wings ; but the
main muscles are only four in number — two elevators and two depressors, which therefore
are the same for both the wings. Among most other insects there are said to be an
elevator and a depressor for each wing ; in the Lepidoptera, Hemiptera, and certain
Hymenoptera (Sawflics) the depressors on each side have coalesced, while in Ants and
their allies the same is also the case with the elevators.

The depressors (PI. XII. fig. 8 & 9, /3) are powerful muscles which occupy a considerable
part of the upper portion of the thorax. They rise from the mesonotum and pass hori-
zontally backwards, lying close to one another along the median line. At their posterior
end they are attached to the two processes of the metaphragma (PL XII. fig. 8) (costal of
Chabrier), an arched process concave in front and convex behind, which, startin<T from
the true hinder edge of the metathorax, passes downwards, terminating in two processes.

The elevators (PL XII. fig. 8 & 9, 6) of the wings lie almost at right angles to the pre-
ceding. They rise from the meso- and metasternum, and passing upwards and forwards
outside the preceding are attached to the wall of the back.

Immediately under the metanotum in this part of the body lies the so-called " meta-
thoracic gland." It consists of a number of large nucleated cells opening into a vestibule
'PL XII. fig. 7) by short minute ducts. The inner wall of the vestibule, at least in the
workers of Lashis flavus, is thrown into several curved ridges, from which proceed a
lumber of strong hairs. The vestibule in this species is elliptic in form and opens to
;he outside by a wide mouth. In other species the shape is different ; in Myrmica
"uglnodls it is somewhat S-shaped and the hairs are smaller; in Lasius fuliginosus it
alls into two divisions, the outer one funnel-shaped, the inner thrown into a number of
spherical chambers. This organ seems to be less highly developed in the males and
emales than in the workers.

* Zeitschr. f. wiss. Zool. 1S7S. f Jahrb. des Yer. fur JSaturkunde im Herz. Nassau, p. 6.

152

SIE JOHN LUBBOCK ON THE ANATOMY OF ANTS.

The abdomen is moved by two muscles (PI. XI. fig. 2, PI. XII. fig. 8) situated in the
so-called first segment or knot.

The first of these muscles occupies the greater portion of the upper part of the knot,
and, passing downwards and backwards, is attached to the lower wall of the abdomen.

The second muscle is attached to the anterior wall of the knot, immediately below the
preceding, and, passing straight backwards, is attached to the upper wall of the
abdomen.

DESCRIPTION OP THE PLATES.

Met. Metathorax.
Sp3. 3rd pair.

The lettering used for the different figures is as follows : —

H. Head. Pr. Prothorax. Mes. Mesothorax.

Sp1. 1st pair of spiracles. Sp". 2nd pair.

Z. Membrane connecting the thorax with head.

.Z1. „ „ the propectus with niedipectus.

Z2. „ ,, the prouotimi with the propectus.

An. Antefurca.

Med. Medifurca.

Po. Postfurca.

gl. Site of thoracic salivary gland.

P. Propectus. Pr, and in some figures B, Pronotum.

A". Processes of the propectus. Gl. 1st ganglion.

G\ 3rd ganglion. G\ 1th ganglion. L1. Base of 1st pair of legs.

D. Base of 2nd pair of legs. V. Base of 3rd pair of legs.

C. Anterior plate of propectus. T. Posterior plate of propectus.

Ab. Abdomen. fm. Femur. tr. Trochanter. Cx. Coxa.

G-

W. Processes of the head.
2nd ganglion.

Muscles.

a. 1st elevator of the head.
b\ 2nd depressor of the head.
c'. 3rd rotator of the head.
e. Protractor of the head.
g. Depressor of antepectus.
r. 3rd muscle of anterior leg.
m. Depressor of prothorax.
/;. '2nd muscle of middle leg.
s. 5th muscle of middle leg.
«'. 2nd elevator of abdomen.
w. 1st rotator of abdomen.
y. 3rd muscle of posterior leg.

2nd elevator of the head. /;. 1st depressor of the head.
1st rotator of the head. c\ 2nd rotator of the head.

& d'. 1th and 5th rotators of the head.

a.
e.
d

f. Elevator of antepectus
h.
k.
n.

1st muscle of anterior leg.
4th muscle of anterior leg.

/'. Elevator of antepectus.
i. 2nd muscle of anterior leg.
5th muscle of anterior leg.

'1-
t.

V.

x.

f-

_i

o.
r.
u.

ft. Depressor of wings.
it. Elevator of abdomen.

Elevator of prothorax.
3rd muscle of middle leg.
6th muscle of middle leg.
1st depressor of abdomen.
1st muscle of posterior leg
lth muscle of posterior leg
(J th muscle of posterior leg.

6. Elevator of wings.

<fr. Depressor of abdomen

1st muscle of middle leg.

4th muscle of middle leg.

1st elevator of abdomen.
v'. 2nd depressor of abdomen.
x'. 2nd muscle of posterior leg.
z. 5th muscle of posterior leg.

SIR JOHN LUBBOCK ON THE ANATOMY OP ANTS. 153

Plate XI.

Fig. 1. Vertical and longitudinal section through the prothorax of Lasius flavus. x 125.

H, posterior wall of head; B, pronotum; P, propectus; Z, membrane connecting the
head and the pronotum; Z\ membrane connecting the head and the propectus; L, base
of leg cut short; An, antefurca. Muscles:— a, first, and a\ second elevator of the head;
b, first, and b\ second depressor of the head ; c, first rotator of the head ; d, fourth, and d\
fifth rotator of the head; e, protractor of the head; /, elevator of antepectus; i, second
muscle of anterior leg ; /, fifth ditto.
Fig. 2. Longitudinal and vertical section through the thorax of Lasius flavus. x 125.

H, posterior part of head; Pr, prothorax ; Mes, mesothorax ; Met, metathorax; C, anterior
plate of propectus; T, posterior plate of propectus; Gl, first, G\ second, G\ third, and G\
fourth thoracic ganglion ; gl, thoracic salvary gland ; Sjr, second, and Sp3, third spiracle ;
Z, membrane connecting the head and the pronotum. Muscles :— a, first, a\ second elevator
of the head ; b, first depressor of the head; c, first, d, fourth, and d\ fifth rotator of the head ;
/, elevator of antepectus ; i, second muscle of anterior leg ; /, fifth muscle of anterior leg ; m,
depressor, and n, elevator of prothorax; q, third, and r, fourth muscle of middle leg; u, first!
and u\ second elevator of abdomen ; v, first depressor of abdomen ; w, first rotator of "abdomen \
x, first muscle of posterior leg.
Fig. 3. Glandular organ at base of prothorax in Myrmica ruginodis. L, upper portion of coxa of

anterior leg. x 200.
Fig. 4. Vertical and longitudinal section through the thorax of Lasius flavus. x 125.

Pr, pronotum; P, propectus; Sp1, first, Sp"-, second, and Sp3, third spiracle; U, base of
anterior leg ; Ls, base of middle leg ; L3, base of posterior leg ; Cx, coxa. Muscles :— c, first, c\
second, and e2, third rotator of the head ; /, elevator of antepectus ; h, first, i, second, k, fourth'
and /, fifth muscle of anterior leg; m, depressor of prothorax; o, 1st muscle of middle leg.
Fig. 5. Thorax of Lasius flavus, seen from above and somewhat flattened out. The external hairs are
omitted, x 125.

Pr, prothorax; Mes, mesothorax; Met, metathorax; Sp1, first, and Sp3, third pair of
spiracles; Z, membrane connecting the thorax with the head; An, antefurca; Sgl, site of the
posthoracic glands. Muscles :— a, first, and a\ second elevator of the head ; c, first rotator of
head; e, protractor of head; /and f, the two elevators of the antepectus; g, depressor of
antepectus ; /, fifth muscle of anterior leg ; m, depressor, and n, elevator of prothorax ; o, first
muscle of the middle leg; u, first elevator of abdomen; w, first rotator of abdomen; x, first,
y, third muscle of posterior leg.
Fig. 6. Longitudinal and horizontal section through the prothorax of Lasius flavus. x 125. Seen from
below.

P, wall of the propectus ; B, wall of the pronotum ; Z\ membrane connecting the pro- and
medipectus ; X, processes of the propectus ; W, processes of the head ; G\ first ganglion ;
An, antefurca. Muscles :— b, first depressor of the head; c\ c2, cut ends of the rotators of
the head ; d, d\ rotators of head ; g, cut ends of depressor of antepectus.
Fig. 7. Section through the prothorax at a rather lower level than the preceding, x 125.

P, wall of the propectus; B, wall of the pronotum ; Z\ membrane connecting the pro- and
medipectus; X, processes of the propectus; W, processes of the head; G1, first ganglion ;
G2, second ganglion ; L, base of the leg ; An, antefurca. Muscles :— c1, second, and c2, third
rotator of the head ; g, depressor of antepectus ; h, first, k, fourth, and /, fifth muscle of anterior
leg (part of these being cut across).

154 SIR JOHN LUBBOCK ON THE ANATOMY OF ANTS.

Plate XII.

Fig. 1. Propectus of Lasius flarus, seen from below, x 100.

H, head ; B, pronotum ; P, propectus ; X, processes of the propectus ; W, processes of
the head ; Z, membrane connecting thorax with head ; Cx, coxa ; tr, trochanter ; fm, femur.
Muscles : — b, first, and bl, second depressor of head ; c\ c-, second and third rotators of head ;
(j, depressor of antepeetus ; h, first, and i, third muscles of anterior leg.
Fig. 2. Longitudinal and horizontal section through the posterior portion of the thorax of Lasius flavus.
x 125.

Pr, posterior margin of prothorax ; Mes, mesothorax ; L-, base of second, and L3, base of
third pair of legs; G2, .second, and G3, third thoracic ganglion; Ab, commencement of
abdomen. Muscles : — u, first, p, second, q, third, r, fourth, s, fifth, and /, sixth muscle of middle
leg ; u, first, and u\ second elevator of abdomen ; v, first depressor of abdomen ; x, first, xl,
second, y, third, yx , fourth, z, fifth, z1, sixth, and r2, seventh muscle of posterior leg.
Fig. 3. Longitudinal and horizontal section through the thorax of Lasius flavus. x 1.25.
Fig. 4. Transverse and vertical section through the prothorax of Lasius flarus. x 125.

B, pronotum; P, propectus; L L, bases of legs; Cx, coxa; G, ganglion; An, antcfurca ; Zi,
membrane connecting the pronotum and propectus. Muscles : — c, first rotator of head; /, ele-
vator of antepeetus; h, first, /', second, /', third, /-, fourth muscle of anterior leg.
Fig. 5. Transverse and vertical section of thorax of Lasius flams passing through the second pair of
spiracles and the base of the middle legs, x 125.

Sjr Sjr, spiracles of second pair ; Med, medifurca ; L L, bases of legs ; n, elevator of thorax ;
o, first, and /, sixth muscle of middle leg.
Fig. 6. Transverse and vertical section through the thorax of Lasius flavus, showing the postfurca. x 125.
Fig. 7. Metathoracic organ of Lasius flavus. x 125.
Fig. 8. Longitudinal and vertical section through the thorax of a queen of Lasius jiavus. x 50.

H, head ; a', elevator of head ; b, first depressor of head ; GG G G, ganglia in depressor of
prothorax ; /3, depressor of wings ; 6. elevator of wings ; tt, elevator of abdomen ; </>, depressor
of abdomen.
Fig. 9. Longitudinal and vertical section through the thorax of a male of Lasius flavus. x 50.
An, antcfurca ; jB, depressor of wings ; 0, elevator of wings.

[ 155 ]

IV. On the Extinct Land- Tortoises of Mauritius and Rodriguez. By Alfred C.

Haddon, B.A., Scholar of Christ's College, and Curator in the Museum of

Zoology and Comparative Anatomy of the University of Cambridge. (Communicated
by Prof. Newton, M.A., F.R.S.)

(Plate XIII.)

Read November 20th, 1879.

JllAVING been recently engaged in determining and cataloguing the large series of
bones of the extinct Land-Tortoises of the Mascarene Islands contained in the Museum
of the University, it has appeared to me that some notes made during that undertaking
might, if published, be of interest to zoologists in general and herpetologists in
particular ; for though a portion of the series was submitted to Dr. Gunther when
preparing his admirable monograph*, yet a very considerable portion, in some instances
containing more perfect specimens, has been received since the publication of that
work, and has therefore never been described, while the whole forms the largest assem-
blage of these rare remains that has yet been collected.

All the specimens in this series were presented to the University by Mr. Edward
Newton, M.A., C.M.G., formerly Colonial Secretary of Mauritius, and now Lieutenant-
Governor of Jamaica. Those from Mauritius were obtained, with one or two exceptions,
from the mud and peat of the Mare aux Songes, where they were found, together with
numerous remains of the Dodo and other extinct members of the fauna of that island,
under circumstances which have been described by the late Mr. George Clark f . The
portion of the series received in 18G7 was examined by Dr. Gunther, and several of its
specimens have been described and figured by him in his work above mentioned ; but
the larger portion did not reach the Museum till the summer of the present year, 1879.
The bones from Rodriguez were received in like manner at two different times ; but they
were all obtained from the caves of that island, the first portion having been collected by
Sergeant Morris, who was working under the direction of Mr. George Jenner, during the
excavations and explorations made at the expense of the British Association for the
Advancement of Science between 1866 and 1871, inclusive $, and the second portion
having been obtained similarly by Mr. James Caldwell on his visit to the island in
1875 §. This last collection was received at our Museum during the past summer and,
like the Mauritian series, which arrived at the same time, has not had the advantage of
Dr. Giinther's inspection.

* ' The Gigantic Land-Tortoisos (Living and Extinct) in the Collection of the British Museum.' By Albert
C. L. G. Giinther, M.A., M.D., Ph.D., F.R.S. London, 1S77.

t l: Account of the late Discovery of Dodos' Romains in the Island of Mauritius. " By George Clark. The Ibis,
1866, pp. 141-140.

t Report of the British Association, 1865, p. xli ; 1866, pp. 401, 402 ; 1867, pp. 287, 288 ; 1872, pp. 23, 24.

§ " Notes on the Zoology of Rodriguez." By J. Caldwell, C.M.Z.S. Proc. Zool. Soc. 1875, pp. 644-647.
SECOND SERIES. — ZOOLOGY, VOL. II. 21

156 ME. A. C. HADDON ON THE EXTINCT

In writing the following remarks, it is needless for me to mention the inestimable
assistance that I have derived from Dr. Giinther's elaborate monograph, as well as from
a comparison of the specimens in our Museum which have been named by him. Indeed
without these latter I could hardly have determined and catalogued our more recent
acquisitions; but since they are so extensive, it seems to me desirable that some
detailed account of them should be given, not with the view of superseding, but of
supplementing, the excellent descriptions which he has already given. I think it cannot
but be gratifying to him to find that the larger amount of materials at my disposal
enables me to confirm in almost every respect the conclusions at which he had arrived.

In all cases I have closely followed his treatment of the subject, in order to facilitate
comparison.

I. Tortoises of Mauritius. Testndo triserrata and T. inepta.

The remains of these, the two common Mauritian species, when sufficiently perfect,
can, with the exception of the femur and the smaller leg-bones, be easily distinguished,
the other forms indicated by Dr. Giintber, Testndo indica and T. leptocnemis, being of
doubtful value. T. indica is known from the shell only, " the upper profile of which is
straight, and not declivous in front;" while T. leptocnemis rests alone on the scapula,
pelvis, and femur ; but its scapula docs not differ from that of T. triserrata, though its
pelvis and femur are " more slender than in either T. triserrata or T. inepta."

Skull. — The additions are : — (1) a portion of the occipital and tympanic regions of a
moderately sized skull ; but not having the maxilla it is impossible to refer it to cither
species.

(2) A slightly imperfect lower jaw of a very large specimen of T. inepta, from the
deposits of Flaeq.

Mandible, extreme length and breadth 7S niillim.

Carapace. — Only a small fragment of a plastron.

Vertebral column. — A single example of the seventh cervical, centrum only. It is
biconcave, the hgemal spine as in T. Vosmari. There is a groove between it and the
tubercles of the anterior zygapophyses, and a deep fossa on the upper surface behind the
anterior zygapophysis on each side, as in the sixth cervical of T. Vosm&ri, and as is also
occasionally found (see examples in the Museum) in the seventh vertebra of the same
animal.

niillim. lnillira.

Extreme length of centrum 38 Length in median line :J4

width „ 27 „ „ 18

„ depth „ 10 „ ., 6

Shoulder-girdle. — Two right and five left can without difficulty be referred to
T. triserrata. Of these there is nothing worthy of notice.

Of the three right and three left which are supposed to belong to T. inepta, two are of
unusually large size. One right shoulder-girdle, marked " '^ bb (PL XIII. fig. 7), has the
coracoid ankylosed : though this character is supposed to be peculiar to T. triserrata,
yet it is found, probably as an individual variation, in this specimen, the compressed

LAND-TOETOISES OF MAURITIUS AND RODRIGUEZ. 157

curved precoracoid (" acromion ") &c. proving it to be T. inepta. This peculiarity can
be parallelled with the case of the corresponding bone of the Rodriguez tortoise.
There is no fresh example of T. inepta, var. Boutoni.

T. inepta. a. b.

millini. millim. millim. millim.

Length of scapula (measured from the suture with coracoid) 177 170

Circumference in its middle 72 72

Longitudinal diameter of glenoid cavity 40 (worn) 40

Length of precoracoid (acromion) 95 92

Length of coracoid . . . . 102 91

Greatest width of coracoid . . . . 69 broken

Least circumference of shaft of coracoid . . . . 49 50

The great amount of variation in the shape of the coracoid in T. triserrata is worthy of
special attention. Though all the well-marked specimens in our collection have passed
through the hands of Dr. Gimther, he only remarks (I. c. p. 16) that " the coracoid lamina
[is] sometimes deeply grooved." The amount of the variations can be well understood
by reference to Plate XIII. figs. 1-5.

The anterior side is sometimes produced into a long spur. The posterior may become
straight rather than bowed for the greater portion of its length, as may also the internal
edge (' epicoracoid ' region). The shaft may be narrow, or broad and flattened; but it
is in the blade that we see the greatest variations. This may be almost smooth, concave
above, and gently convex on its inferior surface, as in the typical example (Giinther,
pi. xxiv. fig. b), or flat superiorly, with the concave inferior surface traversed by prominent
longitudinal ridges, or convex on the upper surface, with the concave lower surface
strongly marked by prominent palmated ridges. The interspaces may become so thin
as to be more or less perforated.

The relative measurements of these bones also vary very considerably, as the following
table shows : —

A (type). B. C. D. E.

millim. millim. millim. millim. millim.

Length of coracoid 74 75 66 76 76

Greatest width of coracoid 61 71 63 58 67

Least circumference of coracoid . . 54 66 47 62 63

It will thus be seen that while the length averages about the same, the width and
circumference of the shaft vary greatly, the normal proportions being, length of coracoid
n, breadth n— 9, circumference n — 15.

The study of all these variations irresistibly calls to mind the remarks of Messrs. A. and
E. Newton*, the truth of which has been subsequently confirmed by the latter gentle-
man and Mr. J. W. Clark f , in their respective memoirs on the osteology of the extinct
Didine bird of liodrignez, as to the wonderful variability of that species. To quote the
former : — " There does not seem to be a single bone in the skeleton of Pezophaps solitaria
which is not liable to greater or less individual variation of some kind or other, . . . but

&

*

" On the Osteology of the Solitaire or Didine Bird of the Island of Rodriguez, Pezophaps solitaria (Gmel.)." By
Alfred Newton, M.A., and Edward Newton, M.A. Phil. Trans. 1SG9, p. 330.

t " On the Osteology of the Solitaire (Pezophaps solitaria, Gmel.)." By Edward Newton, C.M.G., and John
"Willis Clark, M.A., Superintendent of the Museum of Zoology and Comparative Anatomy in the University of
Cambridge. Phil. Trans, vol. 106. p. 451.

21*

158 MK. A. C. HADDON ON THE EXTINCT

the individual variation is not at all confined to absolute size ; it extends to the relative
proportion of divers parts of the bones, to processes or depressions upon thern, such
as are commonly held to be specifically characteristic, so that it is often utterly impos-
sible to predicate any definite limits of individual modification Nor is the

variability of which wc speak wholly dependent on age or sex."

To find that the same results attend the examination of these Tortoise-remains cannot
fail to suggest the possibility that when a sufficient series of specimens of any animal is
compared a similar amount of variability may be observed ; and if this consideration
should lead to nothing else, it should at least make us, in many cases, hesitate about
describing a new species of vertebrate from a single bone, or even from several
fragments.

A second variety of coracoid has been added, which, since it is free, may be safely
referred to T. inepta; of this we have three examples, one, however, being frag-
mentary. They differ from the former type in both form and size (PI. XIII. fig. 6).
The absolute length is much greater than in T. triserrata ; the neck is more slender,
and the posterior edge straighter, as is also the internal margin. Even in these few
examples we see ample evidence of tendency to variation. The great difference of the
length of this bone in the two species is well seen by averaging the relative pro-
portions ; for taking the length of the coracoid to be n, the average breadth would be
about n — 30, and the least circumference n— 47.

Humerus. — Six right and eight left belong to Dr. Gi'mther's group a (T. triserrata).

Two right and four left of T. mepta (group c) ; one of the latter is of enormous size.
In one or two examples the hollow behind the head is extremely shallow, as in
Dr. Gi'mther's group b, thus showing that the latter is merely a variation ; indeed it is
often rather difficult to distinguish between humeri supposed to belong to the two kinds
of tortoises.

2. triserrata. T. inepta.
millim. millim.

Length of humerus from summit of head to middle of trochlea .... 175 250

Narrowest circumference 70 112

Longest diameter of head 43 50

Shortest diameter of head 34 35

Extreme breadth between condyles 71 83

Ulna. — Three right examples apparently belonging to the same species.

millim.

Extreme length 114

Narrowest circumference 44

Radius. — Three specimens, only one of which is perfect. Dr. Gunther's types a
and c may be easily recognized. The one right, a, has a rather slender shaft.

One well-marked right example of c; the other, left, is only a distal fragment; it
would, however, be somewhat larger.

c.
millim.

Length of radius 175

Circumference in narrowest part 38

Longitudinal diameter of semilunar facet 30

Transverse diameter of semilunar facet 20

LAND-TOETOISES OF MAUEITIUS AND EODEIGUEZ. 159

Pelvis. — T. triserrata. We have two pelves nearly perfect, three fragments, and
one right pubis. Two are of great size ; although one of these is of such large size, yet
the os ilium is very narrow and thin. This is probably due to bone-absorption owing to
age. The terminal portion of the pubic process is flattened horizontally, and is sur-
rounded laterally by a deep groove.

The distinctive characters of T. leptocnemis appear of too slight a character to be of
great certainty, especially as no perfect pelvis is known.

T. inepta. Our previous collection of portions of a male and female pelvis has been
greatly enriched by the addition of four perfect pelves and of three halves. There is one
immense os ilium, which is very broad but much compressed, and evidently belonged to
a very old male. There is a femur probably belonging to this pelvis.

T. iriscrruta. T. inepta.

millim. ruillim. millim. millim.
Longest inner vertical diameter of pelvis from summit of ilium

to symphysis 156 . . 138

Longest inner horizontal diameter of pelvis . . 120

Shortest inner horizontal diameter of pelvis hetween ilio-pubic

prominences . . 80

Longest diameter of foramen obfcuratorium . . 36

Width of symphysial bridge . . 25

Breadth of posterior portion of ossa ischii . . 55

Length of os ilii 140 . . 108 147

Least width of os ilii 29 . . 25 40

Proximal width of os ilii . . 54 85+

Femur. — The sixteen new examples have so much the same facies that it is impos-
sible to distinguish between the two species. There does not seem to be a specimen of
T. leptocnemis.

millim.

Length of femur 194

Least circumference 90

Width of condyle 73

This femur probably belongs to T. inepta.

Tibia. — Two right and two left. These show no peculiarity, except that one is
stouter than the others.

millim.

Length of tibia 144

Circumference, least 55

Breadth at upper end 46

II. Testudo Vosmjeri. Rodriguez.

All the remains from this island are referable to one species only, notwithstanding
;he large series of bones examined. The experience of Dr. Gilnther points to the same
conclusion.

Skull. — The additions to our previous fragment consist of one imperfect specimen of
arge size, two occipital crests, and one very perfect example of a young individual, the
ower jaw being also present.

The largest probably measured 11-g- millim. long.

Carapace. — There are two small carapaces, with their plastra, one being very
lerfect, and the posterior marginal plates of a third and large example; also two

160 ME. A. C. HADDON ON THE EXTINCT

plastra of small size, and a portion of a much larger one. In all the thinness of the
bones is very remarkable.

Length of Carapace.

Width of Carapace.

Sternum.

Caudal plate.

In straight line. Over curve.
millim. millim.

In straight line. Over curve.

Length. Width.

Length. Width.

millim. millim.

millim. millim.

millim. millim.

13| i6i

H 17|

G3 73
J4 '1

H H

H+

9

i& 3

2nd.

3rd.

4th.

5th.

6th.

A.

B.

7th.

8th.

lillim.

millim.

millim.

millim.

millim.

millim.

millim.

millim

39

51

66

51

60

64

50

41

10

9

9

10

14

13

18

10

7

8

10

9

16

18

17

18

12

broken

14

20

15

26

40

1

is

18

13

20

12

is

31

Vertebral column. — "We have 74 specimens of cervical vertebras, all being well repre-
sented except the 1st, of which no example is known. The 6th does not present the
foramen noticed by Giinther {lot: tit. p. 57 ; also Ann. & Mag. N. H. 1873, xi. p. 397) ; '
for out of our 13 specimens of this bone only one has a perforation, which is very small,
and might even be accidental. The foramen apparently only occasionally is present, and
that only in old animals.

There are several examples of dorsal and caudal vertebras.

In the majority of cases it has been impossible to distinguish between the slender
and the stout varieties. There is a certain amount of variation in the size of the haemal
spine, in the thickness of the centrum, &c.

Length of centrum

Depth of centrum in the middle ....
Horizontal width of middle of centrum

Width of anterior condyle

Width of anterior glenoid cavity ....

Width of posterior condyle

Width of posterior glenoid

Distance of outer margins of anterior

zygapophyses 16 broken 25 20 29 .. 25 30

Distance of outer margins of posterior

zygapophyses 20 23 23 19 broken . . 2S 37

Shoulder-girdle. — There is great uniformity about these bones, the chief degrees of
variation being in the form of the transverse section of the scapula near the glenoid fossa.
The general form of the precoracoid ("acromion") reminds one of the same bone in
T. triserrata, in being laterally compressed proximally, and vertically compressed distally.
In his monograph, Dr. Giinther remarks (p. 59), " It must be mentioned that there
are in the collection a right and a left scapula, evidently belonging to the same individual,
which differ remarkably from all the others in not having the coracoid ankylosed, although
the individual appears to have been fully adult, and in having the body of the scapula
proper uniformly compressed. It is difficult to account for this apparently individual
aberration." Now, out of the 85 scapulae in our collection, there are 32 in which the
coracoid is ankylosed to the remainder of the shoulder-girdle, leaving 53 in which that
bone is distinct ; of these, 23 are of larger size than the smallest of the ankylosed bones,
several being considerably so. Thus it seems to be a characteristic of this species for
the coracoid to be very irregular in its ankylosis with the coraco-scapula.

It is interesting to note that in one example there is a tendency towards the formation

LAND-TORTOISES OF MAURITIUS AND RODRIGUEZ.

161

of a coracoid foramen by ossification from the coracoid and precoracoid extending into
the epicoracoid region.

Right
Left .,

Total .

Coracoid

ankylosed.

ruillim.

9

23

32

Coracoid free.
Total. (Large examples.)
milliiu. millim.

29

24

53

(11)
(12)

(23)

Total,
millim.

38

47

85

Coracoid fused.

Length of scapula measured from the suture with

the coracoid

Circumference in its middle

Longitudinal diameter of glenoid cavity

Length of coracoid

Greatest width of coracoid

Length of precoracoid (acromion)

A.

millim.

10.5

35

24
49
38

48

B.

millim.

40
30

C (smallest).

luillim.

75

8
19
35

27
39

Coracoid free.

millim.

132 -f

50
304-

50 4-

millirn.

60

48

Humerus. — There is nothing peculiar in our 54 specimens, the largest bone mea-

suring : —

ailliuis.

Length (measured iu a straight line from the summit of the head to the middle of the trochlea ) . . 171

Circumference of the narrowest part of the shaft 62

Longest diameter of head 33

Shortest diameter of head 29

Extreme breadth between the condyles 50 +

Pelvis. — Most of our examples are young, the symphysial bridge still being very
mrrow.

Femur. — One of the 88 femurs is of gigantic size : —

millim.

Tho extreme length being 1G6

The least circumference 68

In conclusion, I would draw the attention of herpetologists to these collections of
VLascarene Tortoise-remains, now in the Zoological Museum at Cambridge, since they
brm by far the most complete series of specimens of these very interesting extinct
■eptiles. I also add my best thanks to Prof. Newton for his kind help whilst preparing
hese notes.

For the convenience of those who may wish to prosecute further the study of these
mimals, as well as to show the amount of materials on which the present remarks are
)ased, I append a list of the specimens in the Museum of the University which I have
consulted.

Testudo triserrata. Mauritius.

6 skulls and 2 lower jaws. (C.) One figured, Giintlier, pi. xxiii. fig. a.

1 skull. New Coll.

4 humeri, right. (C.)

6 „ „ New Coll.

2 „ left (C.)

162 ME. A. C. HADDON ON THE EXTINCT

8 humeri, left. New Coll.
11 shoulder-girdles, right. (C.)

2 „ „ New Coll.

10 „ left. (C.) One figured, Giinther, pi. xxiv. fig. b.

5 „ „ New Coll.

7 pelves (or portions). (C.) One figured, Giinther, pi. xxviii. fig. c; 2 others mentioned.

6 „ „ New Coll.

2 portions of carapaces. (C.)

5 plastrons, adult male. See Giinther, p. 45. One figured, pi. xx. fig. e.
1 carapace. Figured, Giinther, pi. xx. fig. d.

Sp. incert. Mauritius.

1 femur, left. (C.)

8 „ „ New Coll.

6 „ right. (C.)

8 „ „ New Coll.

3 tibia?. New Coll. (One C.)
2

2 radii. „

-*■ y> >>

3 ulnae. „

1 portion of plastron.

1 „ 7th cervical vertebra.

Testudo inepta. Mauritius.

5 skulls and 2 lower jaws. (C.) 1 skull and jaw, figured Giinther, pi. xxiii. fig. b.

1 lower jaw. New Coll.

3 humeri, right. (C.)

2 „ „ New Coll.

3 „ left. (C.)

4 ,, „ New Coll.

1 coraco-scapula, right. (C.)

2 „ „ New Coll.

1 „ left. (C.) T. Boutoni, var.

4 „ „ New Coll.

3 coracoids. New Coll. (2 right and 1 left.)

2 pelves (or portions) . (C.)

8 „ „ New Coll.

Portions of one carapace.

T. friserrata. T. inepta. Sp. incert. Total iu each collection.

Old collection (C.) 48 15 10 73

New collection 28 24 20 78

Total of each species 70 39 36 151

LAND-TORTOISES OF MAURITIUS AND RODRIGUEZ.

163

Testudo Vosmceri. Rodriguez.

Old Collection. New Collection (1879).

2nd cervical vertebra 6

3rd „ „ 7 2

4th „ „ 9 \

5th „ „ f 4

6th „ „ 11 %

7th 12 3

8th „ „ f -

1st dorsal ,, 2

Dorsal „ J

Sacral ,, 3

Caudal ,, ' ' "

Right shoulder-girdle 13

Left „ 6 9

Right coraco-scapula only 14

Left ,, 21 10

Right coracoid 3 1

Left „ 5 Ji

Right humerus 26 fj

Left „ 2U 27

Right radius 20 -

Left „ 17 I

Right ulna 20 5

Left „ 17 b

Pelvis 5 *

Portions of pelvis 33 **

Right femur 21 17

Left „ 18 32

Right tibia 19 *

Left „ 17 12

Right fibula 10

Left „ 7 3

Phalanx terminal 1 . • ■ .

gj-yvj . 1 and lower jaw. 2 and fragments.

Carapace.'.'.'.'!:.';.';...' 2 and 1 fragment.

Plastron

Ribs 3

Total 394 250

y

Total.

8

9
10
12
13
15

7

3

7

3

7

16
15
33
31

4

7
53
47
22
24
25
23

6
66
3S
50
21
29
14
10

1

3

2

2

8

644

DESCRIPTION OF PLATE XIII.

Figs. 1, 2, 3. Under surface of left coracoid of Testudo triserrata (fig. 1, type specimen).
Figs. 4, 5. Under surface of right „ >,

Fig. 6. Upper surface of right coracoid of Testudo inepta.
Fig. 7. Anterior view of right shoulder- girdle of variety of T. inepta.

(All of natural size.)

SECOND SERIES. — ZOOLOGY, VOL. II.

22

C 165 ]

V. On the Morphology of the Skull in the Amphibia Urodela.
By Professor W. K. Parker, F.R.S., F.L.S.

(Plates XIV.-XXI.)

Bead June 19th, 1879.

Species 1. — Salamcmdra maculosa.

1st Stage. — Embryos from the Oviduct a week or two before the time of exclusion * ,
1 inch long, with three pairs of feathery gills.

_L HE embryo of the Spotted Salamander has large external gills, like the larvae of other
" Urodela Caducibranchiata," but in this stage they are at their fullest development.
They soon shrink when the larva is ripe ; in one of these earlier embryos, 1 inch 1 line
long, the left branchiae were just disappearing. These are the youngest I have examined
of this species ; they were not more advanced in development than the larvae of various
congeners only half their size. The chondrocranium gets a considerable start of the
bony deposits, and is rather massive ; yet, although massive, it is a curious piece of
open work (PL XIV. figs. 1, 2), both roof and floor being largely deficient of cartilage.
The cartilage itself is composed of very large cells, and when treated with an ammo-
niacal solution of carmine is an exceedingly beautiful substance, as seen through the
microscope, even with rather low powers.

The skull at this sta°'e has already undergone a considerable amount of modification
by the coalescence of originally unrelated parts, and also by the growth of crests and
conjugational bands on parts that are homologous f . The originally distinct front and
hinder basal bands of cartilage now form one continuous tract on each side ; the foremost
pair, the trabeculae (tr), have united with each both before and behind, and also with
the hinder pair of cartilages — the parachordals, or investing mass {i.v). The latter
cartilages have united near the end of the skull with each other, laterally with the
ear-capsules, and in front, as just stated, with the hind part of the trabeculae, so that
the base of the cranial framework is now well laid down ; it is, however, but two thick
planks of cartilage fixed together, in front and behind, by cross bars in each region.
This is quite in conformity with the Urodelous type of chondrocranium, which is always
largely imperfect in the middle both below and above.

The hinder or parachordal conjugation is a narrow and incomplete band (PL XIV. fig. 2,
i.v, nc), on which the notochord rests ; but the internasal band in front is much larger.

* These and most of my specimens of this Amphibian were the gift of Mr. Tcgetnieier.

t For an account of the earlier conditions of the skull in the Urodela I must refer the reader to my recent memoir
in the ' Philosophical Transactions,' 1877, plates 21-29, pp. 529-597.

SECOND SERIES. — ZOOLOGY, VOL. II. 23

166 PEOF. W. K. PAEKEE ON THE MORPHOLOGY

This new growth of cartilage binds together the primary cornua trabecular between the
nasal sacs (ol, no) ; it is scooped both in front and behind. The scooping in front faces
the re-entering angle between the cornua, and the space there is often a real hole in the
Urodeles — a niesorhinal, vertical passage. The hinder scooping is a correlate of the
swollen fore brain which rests upon it, and from which the olfactory crura are given off,
right and left, to the nasal sacs (figs. 1, 2, c.tr, no). The internasal region of the cornua
is elevated, and on this rest the crescentic olfactory capsules, the simple rudiments of the
nasal labyrinth ; they are now beginning to coalesce with then trabecular undersetters.

Between the eyes the trabecular have sent up a sharp wall to the brain-cavity, right
and left ; regionally, these walls answer to the orbitosphenoids and alisphenoids of the
higher Vertebrata ; and the optic and trigeminal nerves (2, 5) in their exit show the true
landmarks. But in this very primordial and ancient kind of skull the segmentation into
separate neural cartilages is quite suppressed ; only where the notochord lingers can any
trace of somatomic division be seen.

This wall is steeper than it appears to be as seen from above and below (figs. 1, 2, tr)t
but a section (PI. XV. fig. 1, tr) shows its depth. Its inwards curve below makes a floor
to a slight extent ; and near both ends of the wall there is a slight attempt at roofing-in
of the skull (PL XIV. figs. 1, 2). The floor also is slightly chondrified backwards from
the internasal plate. In the young the front margin of this plate (in.c) is a crescentic
notch, finished by the cornua (c. tr).

Behind, also, the floor is more finished ; for the investing mass ensheaths the notochord
near the foramen magnum, and the apices of the trabecular have coalesced in front of its
apex. The perfect arch of cartilage in the occipital region is, at the keystone (s.o),
about one third as broad as the floor ; but the floor is mainly finished by the enclosed
notochord.

In front of that arch there is no cartilage in the roof, the thickest part of the inter-
nasal plate not rising more than halfway to the proper roof.

The occipital condyles (oc.c) are well formed ; they are oval, looking inwards towards
the notochord, and being most seen on the lower side (fig. 2). The huge ear-sacs are
quite perfect as to cartilage ; the curves of the enclosed canals have thrust them into
the sides of the skull, aborting the alisphenoidal crest of the trabecular, so that those end
in front of the junction of the anterior with the posterior canal (fig. 1, a. sc, p.sc) in the
ear-sac. In front and behind the ampullar of the canals make the sac bulbous ; and the
horizontal canal (Ji.sc) forms the eave of the outer face, the rudiment of the tegmen
tympemi. The glossopharyngeal and vagus nerves (9, 10) emerge between the sac and
the occipital ring behind ; and the 5th and 7th (5, 7), or trigeminal and facial nerves,
escape between the sac and the cranial wall in front.

In the fore part of the head the cornua trabecular are squarish ; their inner edge is
thickened whe.ie the simple crescentic nasal cartilage is becoming grafted upon them
(figs. 1, c.tr, ua).

The outer margin of each cornu is cut away in a crescentic manner, and the nasal
sacs lie on and outside the bars ; the outer nostril is near the front laterally, but the
inner (J.n) is further back. Tbis inner nostril is bounded on its inner side by the vomerine

OF THE SKULL IN THE AMPHIBIA UEODELA. 167

tooth-tract (v), and by a small semioval cartilage behind. This is the " ethmo-palatine "
element, a rudimentary preoral bar (fig. 1, behind no). The base of this small, but im-
portant rudiment 19 attached to the most outbent part of the trabecula ; its apex looks
directly outwards.

Besides this antorbital rudiment, there is a large postorbital cartilage in union with
this simple chondrocranium. This is the suspensorium, whose free part is the quadrate,
with its condyle (q). The suspensorium is half the size of the ear-capsule ; it is a trifur-
cate cartilage. The longest and slenderest of the forks is the " ascending process " (a.p) ;
its direction is inwards, upwards, and a little backwards ; it has coalesced with the trabe-
cular crest near its top, and forms a bridge over the first or orbito-nasal branch of the
trigeminal nerve (51).

The next fork is the "otic process" {ot.p); it grows directly backwards, and clings
like a snail's foot to the outside of the ear-sac, embracing, by its pedate end, the ampulla
of the horizontal canal. The third fork or, rather, descending cms is the quadrate region
of the suspensorium (</) ; it grows forwards, so as to end opposite the middle of the inter-
orbital region, and outwards from the skull to a distance equal to the width of the
skull. The oblique end is scooped a little on its upper face for the mandible (fig. 1) ; the
articulation is made at right angles with the long axis of the skull. These are all the
chondrified parts that form this simple skull; yet, few as they are, the elements are
diverse in nature.

The parachordals and the trabecule are most probably homologous tracts ; but the
sense-capsules and the antorbital and postorbital cartilages belong to two other cate-
gories ; the former are paraneurals, and the latter are visceral or facial arches.

Before passing to the lower parts of the face I must remark upon a change that has
begun in the ear-capsule. The walls of this sac had become well chondrified ; but a
remarkable cleft has appeared on its under surface, nearly in the middle. The cleft is
the rudimentary "fenestra ovalis " (fs.o), and the thin ragged edge of cartilage that now
imperfectly hides the enclosed otolith is the rudiment of the stapes, not yet severed from
the capsule *.

The manner in which the fast-growing capsule literally bursts is very remarkable ; the
dehiscence gives rise to a triradiate opening, and in the hinder part of the space there is
a thick ingrowth of the outer wall, which passes into the vestibule upwards and forwards,
and forms a partial irregular septum (fig. 2, fs.o) to the capsule.

As the Urodeles are the fathers of all those that prepare a movable stapes for the more
accurate transmission of the waves of sound, the mode of growth of this part in them is
of the greatest interest.

The lower arches (PI. XIV. fig. 3) are shown in relation ; in the first two the upper
element, or suspensory part, is not shown; that of the first has coalesced with the skull;
that of the second, the hyomandibular, is suppressed in most of the " Caducibranchs " and
in some of the " Perennibranchs," e. g. in Menobranchus. This is a correlate of the early
closing-in of the first cleft ; in no Urodele that I have seen is there the same degree of

* For a Ml description of the formation of the fenestra ovalis and stapes, see my recent memoir on the skull of
the Urodeles in the ' Philosophical Transactions ' for the year 1S77.

23*

168 PEOF. W. K. PAEKEE ON THE MOEPHOLOGT

modification of the " epihyal " element as one sees in most Batrachia. The hasal ele-
ment is deficient in the first two arches, but the second arch is attached to the fore end
of the first basibranched piece ; there are only a first and a second ; and the hyoid and
two first branchial arches are attached to the foremost piece (fig. 3, h.hy, c.br1, c.br2,
b.br1, b.br2).

The free mandible (mk) is a single rod, right and left; it undergoes no subterminal
segmentation. This pair of bars is united together by membrane, so as to form a half-
bent bow. The distal end of each is slender and terete, but further up it gradually
becomes very solid, suddenly narrowing again, to form a rounded angle. Near the angle
above there is a transverse condyloid region, not very distinct from the rest of the car-
tilage. These bars alone have bony films upon them ; they will be described with the
other bony deposits. The length of the mandible (ink) and its place of articulation with
the suspensorium correspond with what we find in the Frog when the gills are wasting.

In the hyoid bar the slender part is above, and the thickening, which is sudden, is
retained, with but little diminution, to the rounded extremity. This bar is Hatter than
the mandible, and has its distal fourth segmented off to form a hypohyal (fig. 3, c.hy, h.hy).

Its size, thickness, and position with relation to the suspensorium correspond with
the stage of the Frog just mentioned.

Each bar passes within, and is attached by ligament to, the mandible ; the distal end
of each hypohyal is loosely tied to the fore end of the first basibranchial (h.hy, b.br1).

Of the next four arches, the branchials, only the two first have a distal or cerato-
branchial element; and the last epibranchial piece (e.brA) does not carry gills. The
first and last have one, and the second and third two rows of denticles forming the gill-
colander ; the four epibranchials lessen from before backwards elegantly. They are all
falciform, but rather thick.

The first basibranchial (fig. 3, b.br1) is evidently the keystone of the two foremost
arches, their ceratobranchials being articulated with it ; it is high, but thickish also,
and its greater height is in front of the first branchial. The second basal piece (b.br2)
is a flattened spatula, whose narrow attached end is articulated to the lower edge of the
first basal segment ; it ends behind in a broad blade, whose margin is straight.

The osseous centres are already numerous ; the largest of these is the parasphenoid,
which forms a thin floor of lath-like bone to the base of the chondrocranium from end
to end, or nearly. The apex of the notochord is ensheathed by an imperfect " cephalo-
style;" but this is now wasting, and is not distinct from the parasphenoid: in some
Caducibranchs (e. g. Seironota) this centre ensheaths the fore half of the notochord and
is autogenous.

Above, the frontals and parietals are forming the roof ; these will be described in the
next stage ; the squamosals (sq) are applied as splints to the suspensoria.

Distinct premaxillaries, dentaries, splenials, and articulars (px, d, sp, ar) are now well
seen ; the maxillaries (mx) are later in their appearance than these.

The fore part of the palate is beset with teeth, like the pharyngeal teeth of Osseous
Fishes ; these are arranged in an arc on each side, whose concavity is inwards. The
fore part of each tract is broad, and in this part the teeth are cemented to a thin bony

OF THE SKULL IN THE AMPHIBIA TJRODELA. 169

plate, the vomer (») ; the rest thin out into a small row of denticles, only attached to
bone in front. The bone whose fore end carries these hinder teeth is the pterygopalatine
(p.pg) ; it commenced as a palatine, carrying its own share of the teeth, but threw out
a pterygoid spur, which the teeth did not follow in its diverging course. The teeth
will approach still nearer to the mid line, and the non-dentigerous plate will become
segmented from the rest, and bend outwards to the zygoma.

The relation of the cartilaginous to the bony parts will be still better seen by reference
to the figures of sections, which were made from an embryo somewhat more advanced.

The first transversely vertical section (fig. 4) was made just in front of the internasal
plate (iu.c) ; it shows the thickness of that plate, and the position of the nasal roofs (ol)
and trabecular cornua (c.tr). These latter make a floor quite similar to the roof; the
roof itself has already been grafted on to the internasal plate. Where the union is
taking place there the nasal processes of the premaxillaries (n.px) are seen in section,
whilst below, the vomer and its teeth (o) are seen beneath the trabecular nasal floor.

The second section (fig. 5) shows the same bones and cartilages as the first, but at the
hinder margin of the internasal plate (iu.c). The olfactory nerves (1) find their way to
the nasal sacs between the upper and lower cartilages (ol, c.tr), and the middle space is
continuous with the cranial cavity.

The third section (PL XV. fig. 1) was made directly through the eyeballs (<?), through
the middle of the brain (lettered c2 by mistake).

The trabecular (tr) have become strongly crested, and the crest at this part is orbito-
sphenoidal; the floor is finished by the gently bulging parasphenoid (pa.s), and the roof
is filling in from the edges by the frontals (f).

On the palate, one third the way from the skull to the face, the pterygoid spur (pg)
has been severed. The rest of the skeleton here to be seen is below the mouth. The
outermost of these bars is the mandible (mk), a ad its ovoidal section shows the articulare
on the inside and the dentary on the outside (ar,d).

Next to these come the ceratohyals (c.luj) ; they have an ovoidal outline, but the
narrow end is below, whilst in the mandible it is above. The ceratobranchials are severed
close behind the first basibranchial, so that the fore end of the second is cut through
(b.br-) ; the lesser second ceratobranchials lie in a higher plane than the larger first pair
(c.br1, c.br'2).

The fourth section is a little oblique (PL XV. fig. 2) ; it is behind the eye, and catches
the mid brain (c2) and backwardly turned lower part of the fore brain ; the top of the
infundibulum (inf) is seen in it, but not its connexion with the pituitary body.

On one side the fore face of the ear-sac (an) is laid bare, but not on the other ; on this
latter side the pedicle of the suspensorium (jpd) is seen to ascend over the orbito-nasal
nerve (51), and to be coalesced with the alisphenoidal crest of the trabecula (tr) ; there
is no union besides this ascending process to the skull wall. The quadrate cartilage is
in relation with the articular end of the mandible (q, ar) ; from its upper edge the otic
process (ot.p) is seen to rise.

On the other side the ascending process is cut through, and the Casserian ganglion
is exposed; it is seen between the trabecula (//') and auditory sac (cm). The para-

170 PROF. W. K. PARKER ON THE MORPHOLOGY

sphenoid lies between the trabecular sections, and the parietals are finishing the roof
over the hind brain. The sections of the arches are made through a more backward
part ; they differ but little from the last, yet the hyoid (c.hy) has been severed where
it is smaller, and the second basibranchial lies low in the throat ; the pterygoid bone is
running towards the quadrate (q).

The fifth section (PL XIV. fig. G) is through the apex of the notochord (nc) ; the
anterior semicircular canal (a.sc) is severed near its junction with the posterior and the
horizontal canal (h.sc) through its middle.

The roof of the ear-capsule is perfect, but the floor is imperfect ; for the fissure for the
fenestra ovalis is now gaping, and part of the jagged edge of cartilage is becoming
rounded into the stapes (see PI. XIV. fig. 2,fs.o).

The superoccipital margin is imperfect here ; below, the imperfect cephalostyle (nc) is
seen to be united with the parasphenoid.

The sixth section (Plate XV. fig. 3) was made behind the superoccipital roof, through
the posterior semicircular canal and the hind part of the horizontal (p.sc, h.sc). The
cartilage grows inwards round the latter, and mesiad of this inflection we see the cleft
which becomes the fenestra ovalis. This section is behind the occipital roof, and through
the middle of the cranial notochord and the most solid part of the basal plate.

The seventh section (Plate XV. fig. 4) is near the back of the ear-capsule ; it is the front
face of a solid section that is figured, taken a little in front of the condyles for the atlas.
The ear-walls are seen to be thick, and the sides of the parachordals (iv) are solid as com-
pared with the conjugational plate on which the huge notochord lies. The razor has passed
through the glossopharyngeal and vagus nerves (9, 10) at their exit, the former of these
nerves being close to the membranous labyrinth, and burrowing the periotic cartilage.

2nd Stage. — Ripe embryos of Salamandra maculosa, 1§ inch long.
(" Cryptobranch stage.")

Like the young of the Surinam toad (Pipa), the young Salamanders at birth are ready
for terrestrial life ; they, however, have had branchice before their escape from the oviduct.
The young Pipes never develop gills, and are developed in quasi-uterine pouches. The
changes seen in the skull of this stage as compared with the last are very great ; they
probably only took ten or twelve days for their development : the chondrocranium, in
the earlier stage, was very perfect, ripe for metamorphosis into a true Salamandrian
bony skull.

The basal region (PL XV. fig. 6) is large and oblong ; and if the suspensoria be
removed, the sensory organs, in front and behind, form nearly equal expansions. In
the earlier stage the hinder, or auditory, were very large compared with the front, or
olfactory masses.

Prom the foramen magnum to near the front of the internasal plate the whole base is
floored by one huge lamina of bone, the parasphenoid (pa.s), and through it the nakedness
of the proper cranium can be seen.

Prom the region between the internal nostrils (i.n) to that between the foramina
ovalia the whole floor is imperfect, so that there is scarcely a trace of cartilage

OF THE SKULL IN THE AMPHIBIA UEODELA. 171

beneath the fore and the hind brain. This space is called pituitary ; but the pituitary
body only occupies a foot breadth at the end of this great field. This main basal
fontanelle is followed by another, the " posterior basicranial fontanelle " of Rathke.
This is formed by a reopening of the original membranous space on each side of the
notochord, between the ends of the trabecular and the apices or fore ends of the para-
chordals.

The cranial notochord is still large; it is bounded in front by the trabecular, and
behind it lies on the basal plate ; it fills the space between the occipital condyles, as it
emerges to become the axis of the spinal column (fig. 2, nc, oc.c).

The occipital arch has become bony, all but the mid line, above and below (e.o,f.m).
This bony tract stretches over the epiotic and opisthotic region of the auditory capsule.

Scarcely distinct from this tract below is another (pro) which runs in a crescentic
manner as two wings of bone, one on the antero-inferior, and the other on the antero-
superior face of the auditory capsule. Thus the " sacculus " of the fenestrate region is
still floored with cartilage : a band of cartilage also intervenes between the epiotic and
prootic bony tracts, and along the actual front face of the capsule ; this latter tract has
a plaster of cartilage upon it, derived from the basal plate, and this generally remains
soft in relation to the suspensorial pedicle. The fenestra ovalis is now perfect, and is filled
with a thickish cartilaginous stapes (fig. 6, st). The skull in front of the car-sacs is
entirely unossified, yet there has appeared a bony plate on each side, which in osseous
fishes is an ectostosis. This is the prefrontal or ectethmoid (e.eth).

The sphenoidal walls of the trabecular (tr) are more perfect, but scarcely overlap
above, so that the fontanelle is complete. In front the nasal roofs (na) are of great
extent, and they are quite confluent with the very wide trabecular cornua. The ethmo-
palatine cartilages (fig. 6, outside p>a) have also united themselves to the posterior edge
of the nasal curtain.

Other most instructive characters have now appeared : the internasal plate (in.c) has
lost its hinder projection and has gained one in front ; this is the prenasal rostrum (p.n),
a process not common in the Urodeles, and suppressed in many of the Anura. In the
latter group the primary cornua trabecular give off a small process, the " prorhinal ;" it
is a secondary cornu : this is fairly seen in the ripe young of the Salamander (fig. 6, c.tr),
and will persist in the adult (PI. XVI.). The rest of each cornu forms the subnasal
lamina, or trabecular floor of the nose, as in the Frog.

The large suspensorial cartilage is still only confluent with the skull by its ascending
process (a.p) over the first branch of the fifth nerve; the " pedicle " proper is never more
than a bulging of the lower face of the ascending process in this species. The otic
process (otp) is well developed, and clings by its expanded face to the most projecting
part of the ear-sac. A pterygoid process (e.pg) has appeared.

The quadrate region of the suspensorium is now largely ossified (q), as is the rule in
the tailed Amphibia, and the exception in the Anura ; the condyle still looks forwards,
and reaches to an imaginary line across the middle of the skull.

Thus we see that, including the quadrate bones, there arc only three pairs of osseous
centres found, as yet, in the substance of the chondrocranium.

172 PEOF. W. K. PAEKEE ON THE MOEPHOLOGY

Many other bony plates, however, have been developed in the overlying fibrous tracts ;
the chief of these is the parasphenoid (pa.s). This bone rounds gently to a blunt point
behind, beneath the foramen magnum, and to two blunt points in front, close behind
the prenasal rostrum ; the basitemporal region is widest and is notched. Running in an
arched manner round the front of the parasphenoid is a chain of four dentigerous bones ;
the foremost of these are the larger, these are the vomers («) ; they keep their width
throughout. Behind the vomers are the palatines (pa) ; these are little wedges, with
their point behind. In the last stage they were continuous with a pterygoid expansion ;
this is now a large distinct bone.

This pterygoid bone (pg) is a large broad wedge, whose point, severed from the pointed
end of the palatine, looks forward, but lies further outwards than the end of the palatine.
Its hinder end is roughly semicircular, and forms a clamp to the suspensorium from the
exit of the trigeminal nerve inwards to the quadrate condyle outwards. The outer
surface of the suspensorium is bouDd by its own splint, the squamosal (sq), which also
helps to make the tegmen tympani outside the horizontal canal.

The great upper fontanelle is closed in by two pairs of bones, the frontals and parietals
(f, p), that are nearly equal ; the parietal is a much larger bone in the Urodeles than
in the Anura.

The parietals not only run in and form a wall-plate to the trabecular crest, but they
also overlie the hind part of the frontals ; these, again, are in their turn overlapped by
the nasals and nasal processes of the premaxillaries. The nasals (fig. 5, n) are small
subcrescentic bones ; they cover in but little of the nasal roof. The nasal processes of
the premaxillaries [n,px) are as long as, but narrower than, the marginal part of these
bones : they are larger in the Urodeles than in the Anura. External to these the
maxillaries (mx) have enlarged ; they have an ascending facial part, and a styloid
zygomatic process (fig. 5) ; but the palatal part, as in the premaxillaries, is very narrow.
On their upper edge, below the external nostril (e.n), there is a small septo-maxillary
(s.m.v).

Between the hind part of each maxillary and nasal, on the hinder part of the nasal
dome, there is a thin bony lamina, answering to the ectosteal plate of the Fishes' pre-
frontal (ectethmoid) ; it retains its distinctness here, and thus has to be called the
external ectethmoid (e.eth). The sphenethmoidal bony centres are all that are yet
wanting to make up the sum of the bony centres seen in the skull of the adult.

The metamorphosis of the inferior arches takes place after birth ; they have now
merely increased in size ; the branchials are now at their fullest development, and will
soon suffer extensive absorption.

3rd Stage. — Skull of the adult Salamandra maculosa.

The skull of the adult of this species shows at once the likeness and the unlikeness of
this type to that of an average Batrachian ; yet in reality this skull differs as much
from that of Rana temporaria or of Btifo vulgaris as it did in its early conditions and in
the mode of its metamorphosis. But scattered up and down the great Batrachian
" Order " there are remarkably generalized types, whose skulls, now in this and now in

OF THE SKULL IN THE AMPHIBIA URODELA. 173

that, agree with this of the highest kind of " Caducibranchiate " Urodele. I shall
therefore carefully set down the characteristics of this skull, and then it can be used as a
measure, being " perfectissimum in suo genere," of the lowness or height, in type, of the
skulls of other Urodeles, and also as a test of Avhat is normal or aberrant in the skulls of
the Batrachia. This sj)ecies and Ran a temporaria, therefore, will be taken as con-
venient, and yet worthy, representative Salamandrian and Batrachian types. They are
culminatiug forms, being most perfectly specialized according to then- kind. The
general resemblance of this skull to that of a Grey Frog is seen at once (PI. XVI.
figs. 1-3) ; but attention to details will give us a large number of important differences.

When this skull is stripped of its investing bones (fig. 1) we see how little change,
except that of ossification and increase of size, has taken place. The very strong occipital
arch has this normal amphibian weakness, namely, that an oblong tract both of roof
and floor remains unossified ; in the floor the shrunken notochord still remains.

The condyles (oc.c) are best seen from below (fig. 2) ; they arc large, oval, and slightly
pedunculated ; they are very wide apart. A large rounded fossa divides this arch above,
on each side, from the periotic mass, but there is no suture ; below, the regions melt
insensibly into one another (fig. 2). But both below and on the inside (fig. 5) the double
passage for the glossopharyngeal and vagus marks the bounds of these two territories.
The occipital ring expands greatly on each side to embrace and unite with the large ear-
masses ; but both above and below the shortness of the cartilaginous tracts causes an
emarginatiou of great size, both before and behind, in this annular growth.

The front emarginatiou of the floor is the hinder margin of the posterior basicranial
fontanelle; its fore boundary is formed by a bridge (a.tr.), which is cartilaginous
in the middle and bony at its large piers. Did these osseous tracts meet, we should have
the " prootic bridge" of the Bony Pishes; the real morphological nature of this band
is of great interest.

The hinder part or threshold of the skull was formed by the separate and somewhat
late " parachordals," or investing mass ; the bridge is formed by the posterior ends of
the trabecular, which had united, in the first stage described, with the parachordals, and
ran into each other in front of the notochord (PI. XIV. fig. 1, nc, tr, iv). So that two
things have taken place — much of the cartilage has become bone, and old cracks have
opened again, by relative shrinking of that which does not ossify.

The occipital arch is not more strongly united to the auditory masses than the
trabecular walls ; the alisphenoidal region is strongly cemented to these masses by a
remarkable trespass of each prootic centre.

The early condition of the trabecula alone can explain this curious encloskeletal skull.
These rods embrace the notochord by their hinder ends, and then turn rather suddenly
outwards (Huxley on Menobranchus, P. Z. S. 1874, pi. xxxi. figs. 1, 2, Tr, Gh).

In this adult skull the confluent ends of the trabecular form a thin narrow band of
cartilage, passing transversely (with a little backward deflection) across the cranial
floor. Towards the side each bar suddenly thickens, and is bent on itself so as to form
an acute angle ; in front of the bend it becomes very solid, and at the bend there is the

SECOND SERIES. — ZOOLOGY, VOL. II. 24

174 PEOF. W. K. PAEKER ON THE MOEPHOLOGY

alispkenoidal crest. Between this crest and the ear-mass the great trigeminal nerve
passes out through the foramen ovale (5).

The synchondrosis seen in front of the bend does separate the orbito-sphenoidal {sp.e)
from the alisphenoidal regions ; hut the bony matter behiud the cartilage is all an
outgrowth from the prootic. This outgrowth is double, growing, as we saw, into both
wall and floor. The bony matter is not perfect in the inside (fig. 5) ; there is a
vertical band of cartilage directly in front of the foramen ovale (5), which broadens both
above and below.

The optic foramen (2) is formed in the middle of a large membranous fontanelle, as in
the Frog; this space is margined with cartilage on the outside (fig. 3. 2). In front of
that ring there is solid bone up to the olfactory capsules (na) ; those right and left bony
walls correspond to the annular sphenethmoid of the Frog, but they are not conjugated
together, either above or below. The coping of this hardened wall is thick (fig. 4, sp.e),
but below it thins out and turns inwards. There is some floor to the skull iu front ; for
the internasal plate (i.n.c) is scooped, and the fore end of each bony wall is cemented to
the other by an elegant crescentic growth of cartilage, which has a thickened rim.

The crescentic edge of the internasal plate below is of much smaller extent than the
upper, and extends further back; therefore there is some floor; on each side, just in
front of the bone, the olfactory crus escapes (figs. 4 & 5. 1). The large internasal plate
takes in most of the cornua below ; but the internal angle of each is developed into an
inturned tongue of cartilage, the " prorhinal " (c.tr) ; and between these two a similar
process grows from the mid line : this is the prenasal rostrum (p.n). These three out-
growths seem to bo the non-segmented rudiments of the paired and unpaired elements
of the foremost visceral arch, whose splints are the premaxillaries.

The thickened internasal plate above is shaped like the letter X ; for its hind margin
is crescentic, its fore margin deeply notched in a rounded manner, whilst its solid middle
part is bounded by a crescentic ridge behind, and in front grows into diverging ridges,
the ridges to which the backs of the olfactory crescents are cemented. This structure
is eminently Selachian, but also foreshadows the nasal labyrinth of the higher Ver-
tebrata.

The original crescent of cartilage closes round the outer nostril {e.n), and this is
further occluded by a membranous valve, which is fan-shaped, with the narrow part
looking; forwards. The nasal sacs are hollow below ; but the trabecular floor them to some
extent on the inner side. This part is invested by the huge vomers (v) ; but the ant-
orbital part, besides dipping down, has a new selvedge added to it. This is the cthmo-
palatine cartilage {e.pa), which is very apt to coalesce with the antorbital part of the
nasal capsule.

The auditory capsule is nearly all solid bone ; but there is a cartilaginous tract outside
the ampullae of the anterior and horizontal canals (a.sc, li.se). This is a character of
considerable morphological import ; for this unossified tract was originally a growth from
the basal plate, and with this the otic process of the suspensorium coalesces.

There is a well-ossified lip, like the mouth of a pitcher, to the fenestra ovalis : it looks

OF THE SKULL IN THE AMPHIBIA UKODELA. 175

outwards and backwards ; the thick, oval, closely-fitting stapes (st) remains cartilaginous,
and is only attached to the suspensorium by a ligament *.

The suspensoriuin has changed greatly since the time of birth ; in direction it has
moved its condyle from being opposite to the middle of the skull to an imaginary line
drawn across at the end of the paraspbenoid : this is equal to its average position in
adult Batrachia. It is foot-shaped below, the base looking outwards and backwards ;
all but the articular face is ossified (q) up to the middle of the otic process ; also the part to
which it adheres, and with which it becomes confluent, is soft. The ascending process
(pel) is small and terete ; it remains confluent with the alisphenoidal region. This
pedicle has a bulbous enlargement, like that which bears the condyle of the pedicle
in a Frog.

But little of this expansion, which articulates with the ear-mass, is seen from below
(fig. 2), for it is ensheathed by the pterygoid {pg). This bone is hammer-shaped, and its
long head binds on the pedicle within and on the inside of the quadrate by its outer
lobe. The handle has used up the cartilage that grew from the suspensorium (PI. XV.
fig. 5) ; it is grooved where the cartilage lay (figs. 1, 3, 4), and its bluntly pointed end
has turned outwards to be tied by ligament to the zygomatic process of the maxillary
(mx).

On the other hand, the bone from which it was segmented, the palatine (fig. 2, pa),
has grown further inwards, and reaches further backwards than the foramen ovale (5).
It has coalesced with the dentigerous region of the vomer (v), and the two tooth-
bearing tracts are rib-like bars of bone, whose arrangement under the parasphenoid is
lyriform. To one fresh from mere mammalian morphology, these sigmoid rods, and the
edentulous plates growing like wings from them in front, would be an inexplicable
puzzle.

The anterior expansions are a new development of bone from the simple larval
vomers, and are a caducibrancliiate character. They are very elegant ; between these
wings there is a large semielliptical notch revealing the internasal cartilage and
rostrum ; their margin at this point is curled upwards. Their outer margin is lessened by
a large rounded notch, which bounds the internal nostril (i.n) ; behind this there is a lobe
which does not extend so far outwards as the fore part. They unite by " harmony " for a
short distance between the dentigerous bars ; the internal nostril has its rim finished,
outside by the maxillary (mx), and behind by the ethmopalatine cartilage (e.pa).

The maxillaries have much, and the premaxillaries some, palatine expansion (fig. 2,
px, mx) ; altogether this is a very hard and finished palate for so low a type. The
palatine process of the premaxillary is short (fig. 2) ; its nasal process (n.p.i) is long and
inbent towards its fellow at the middle ; the frontals run under these bars, and in front
of the frontals the internasal cartilage is seen. The maxillaries (figs. 1 & 3, mx) have a
large and swollen facial plate ; and behind the nasal region these bones run back

* Mr. A. Doran lias shown me a stapes, said to belong to this species, which is ossified and has a stalk. I find
nothing of the kind in tho specimens dissected by me ; and I think it probable that that specimen (in the Hunterian
Museum) belonged to another kind. Hereafter I shall have to describe this sort of stapes in Spelerpes and
Desmo'jnatlms ; it exists also in the Menopome.

24*

170 PEOF. W. K. PAEKEE ON THE MOEPHOLOGY

beyond the apices of the pterygoids as far as in an average Batrachian : there is, however,
no quadrato-jngal to bind them to the quadrate ; they are free as in Osseous Fishes. The
dentigerous line, continuous in front with that of the premaxillaries (fig. 2), extends
nearly to the distal end. Above, at tbe junction of these bones, and outside the external
nostril (e.n), there is a small graniform septo-maxillary (figs. 1 & 3, s.mx). The squa-
mosals (figs. 1 & 3, sq) have a convex, suboval supratemporal region (fig. 1), and a large
oblong preopercular part (fig. 3) ; above, they lie over the two first ampullae and the
tegmen tympani, whilst the descending part strongly binds upon the outer face of the
suspensorium. After being broad above, and narrowing somewhat down the shaft, the
bone has a broad arcuate base fitting against the expansion of the quadrate, towards the
condyle.

Over the nasal region there are two pairs of bones on each side ; the nasals (n) are the
inner, and the ectethmoid plates (e.eth) the outer of these. The nasals have an anterior
margin, which is concave, to fit against the external nostril (e.n) ; their main part is
wedged in between the nasal processes of the premaxillaries, the ascending plate of the
maxillaries, and the ectethmoids ; their hinder angle rests upon their frontals.

These latter bones (e.eth) are oblong ; they form, as it were, two horns to the pair of
frontals (f), stretching outwards and forwards ; the orbito-nasal nerve (51) passes into the
olfactory sac between their outer end and the high part of the maxillary. These bones
remain distinct. Taken together, the frontals and parietals (fig. l,f,p) have an hour-
glass-shaped outline ; this figure is made up of four nearly equal bones.

In front, the frontals are overlapped by the nasals, nasal processes of the premaxillaries,
and ectethmoids ; behind, they repeat this imbrication by lying on the parietals, towards
which they narrow, whilst towards them the parietals are narrowed. The superorbital
margin (figs. 1 & 3) is strong, but not overarching ; there the parietal runs forward,
coping the cranial wall.

Behind, the parietals are trilobate ; the first or outer lobe binds on the front of the
ear-mass; the middle lobe covers the junction of the anterior and posterior canals; the
third or inner lobe meets its fellow over the supraoccipital region, tiling that roof nearly
to the edge of the foramen magnum.

The great parasphenoid (figs. 2, 3, 5,pa.s) has its narrow anterior part invested by the
vomers, and its middle region by the long palatines ; behind, it is wide, and rendered
irregularly crenate by four pairs of notches; there are three denticulations on this
foliaceous part.

The inferior arches now come under review. The first of these is the free mandible ;
this is now a flat arcuate bar, well ossified by the long sheathing articulare (figs. 3 & 5,
ar), which has ossified nearly all the cartilage except the oblong cylindroidal condyle.
The articulare reaches far forwards, but is encased in two splints ; in front of the
condyle it sends upwards a coronoid process. The external dentigerous splint, the
dentary (d), reaches from the chin to the angle of the jaw, and ends only slightly in front
of the end of the articulare. On the inside it lies over the edges of the jaw, but is
deficient in the middle. This deficiency is made up by a long curved dentigerous
lath, the splenial (fig. 5, sj)) ; the bone is pointed at both ends, and reaches from the

OF THE SKULL IN THE AMPHIBIA URODELA. 177

chin to the coronoid process. The Salamander has no inento-Meckelian hone, and the
Frog has no splenial. The angle of the jaw is attached to the outside of the apes of the
ceratohyal hy the mandibulo-hyoid ligament {c.hy, m.h.l). The same part of the hyoid
is attached to the suspeusorium by the hyosuspeusorial ligament.

The suspensorium is attached to the oval cartilaginous stapes by the suspensorio-
stapedial ligament. The first cleft closes early. Thus there is no tympano-eustachian
cavity, no " annulus," and therefore no membrana tympani. The columella is re-
presented by the suspensorio-stapedial ligament, over which the facial nerve passes.

Thus the high type of Urodele is much less specialized in these respects than the Frog.
It is a much higher creature than the Frog's Tadpole prior to the absorption of the tail,
for in its own diverging line it ascends to a considerable vertebrate height ; yet in some
metamorphic peculiarities it rises no higher than the Batrachian larva.

The hyobranchial series of arches have undergone great change (PL XIV. figs. 3 & 7),
and have all continued soft except the distal part of the 2nd basibranchial (b.br).

The hyoid bar {c.hy) has become much flattened, and its ventral end is the wider ; the
distal piece is quite detached, and is now directed forwards instead of backwards, parallel
with the 1st ceratobranchial (c.br1), which is a longer rod and is more directly
attached to the ceratohyal.

The 1st epibranchial (e.br1) has reached the median piece, some distance behind its
distal segment. It is still a large cartilage, but is made still larger by the confluence
with it of the 2nd epibranchial, which is now fused with its own ceratobranchial piece
(fig. 3, c.br). These latter bars, right and left, articulate with the oblique end of
the 1st basibranchial (b.br1), which now has four cartilages directly articulated to it
instead of two (figs. 3 & 7). All except the broad end of the 2nd basibranchial (b.br)
has been absorbed, and that has become a F"-shaped bone, with the tips of its crura soft.
It is now a median bilobate " thyrohyal," being applied to the front of the larynx —
a very simple structure, but cartilaginous.

This is, on the whole, the kind and degree of modification undergone by this series of
cartilages in the " Caducibranchs ;" but we shall find some curious modifications in
other types of the group.

Species 2. — Nbtojahthalmus viridescens.

1st Stage. — Larva? 1 inch 2§ lines long *.

My large larvae of this North- American Newt show some very instructive conditions of
the urodelous skull ; the skull of the adult also is very important. The larvae at this
stage are as much developed, on the whole, as the ripe young of the Spotted Salamander
(PI. XV.) ; I suspect them to have been nearly ready for metamorphosis into Caduci-
branchs. They show a remarkable difference as to the time at which certain changes
take place. In some respects they are far less developed than my first stage of the type
just described (PI. XIV.), and in others are far more advanced than the second. The

* This and the remainder of the types to be described in the present paper are the gift of Prof. St. George
llivart, F.P.S.

178 PEOP. W. K. PAEKEE ON THE MOEPHOLOGY

adult skulls have much in common ; but this is move massive (Pis. XVI. & XVII.) and
longer.

In these full-gilled larvae the occipital arch (PI. XVII. figs. 1-3) is well ossified, and
the ossification on each side runs continuously over the inner part of each auditory sac,
so that the exoccipitals and prootics (e.o,pro) are not separate. There is a distinct and
widish tract of super- and basioccipital cartilage (figs. 1, 2), and a solid facet of cartilage
for each condyle (oc.c). These oval elevations look backwards and inwards. The inter-
space between them is square and large ; it is filled by a rudimentary vertebra, whose
centrum has coalesced with the centrum of the " atlas," between its large pedunculated
" prsezygapopbyses."

In one specimen (figs. 1 & 2) there was scarcely more conjugating cartilage in the basi-
occipital region than is seen in the adult Proteus and Menobranchus s, for the parachordals
(iv) had retreated already from the trabec ulae, whose hind part (its parachordal tract)
had been already absorbed (figs. 1, 2, tr, iv, nc) ; but in the other larva (of the same
size) there was a U-shaped posterior basicranial fontanelle, traversed by the thick
notochord, and bounded in front by the thick broadish trabecular band (fig. 3, tr, nc,
p.bc.f, iv). This is the band which is persistent in the Salamander (PI. XVI. fig. 4), but
which so soon vanishes in this type, but not always at the same time in different
individuals.

Laterally the trabecular were confluent with the parachordal cartilages and also with
the cartilaginous regions of the ear-sacs ; beyond that point they had grown into side
walls to the skull, which were of great height in reality, but bulged considerably.

In the ripe Salamander's larvae (PI. XV.) there was no appearance of bony deposit in
the side walls ; but here, from the optic foramen (2) to a short distance from the nasal
region, the whole wall is solid bone (figs. 1 & 2, sp.e). Yet in front of these tracts the
whole of the structures represent a very early condition in other types (e. g. Siredon,
Salamandra).

Three pairs of splints have not yet appeared, namely the nasals, ectoethmoid, and
maxillaries ; this stage thus corresponds with the adult Proteus and Menobranchus *.

Three pairs of cartilaginous growths are seen here, of which only one pair occurs in
the first of those " Perennibranchs ;" and one tract seen in them is wanting here ; for the
long internasal part of the trabecular has no conjugatory tract, as in Proteus and
Menobranchus', but what they lack these have, namely, a large flat lobe outside the
end of each rod (c.tr) ; these give the form of a hatchet to the fore part of the trabecular.

These larvae correspond with those Perennibranchs in having a distinct " antorbital "
or ethmopalatine cartilage (e.pa) on each side ; it is, however, shorter and broader than in
them (Huxley, op. cit. pi. xxx. fig. 1, A.o, and PL XVII. figs. 1 & 2, e.pa). But they
have acquired a new structure of intense importance ; this is a sickle-shaped olfactory
cartilage, or nasal roof (na) f , which is absent in Proteus and imperfect in Menobranchus.

* For the figures of these types, see Huxley, P. Z. S. 1874, pis. xxix. xxx., and my recent paper in the Phil.
Trans. " On the Skull of the Urodeles."

t After Prof. Huxley's paper was published, he expressed to me his doubt of the distinctness, in any case, of this,
the foremost of the " paraneural " cartilages : the figures here given may serve to expel that doubt. The chondro-
cranium of Menobranchus, as figured by him, has much less basal cartilage in it than it had at an earlier period.

OF THE SKULL IN THE AMPHIBIA UEODELA. 179

In Salamandra the lobe sent outwards from each trabecula forms the subnasal lamina,
and the end of the trabecula gives off in the adult the " prorhinal " process.

The auditory capsules are still very large, relatively ; above, the anterior and posterior
canals (a.sc, p.sc) are surrounded by bone; and below (fig. 2) the vestibule is sur-
rounded on its inner side by a somewhat narrower tract. The huge semicircular canals
above (fig. 1), and the swelling vestibule with its enclosed otolith, give form to the
auditory sac ; the horizontal canal (h.sc) affects the form most, bulging over the side
and enlarging the fore part so as to make the sac pyriform. The stapes (fig. 2, st) lies
behind the middle ; it is an elegantly oval plate, and lies in a fenestra ovalis much too
large for it. Between the ear-sac and the occipital ring the glosso-pharyngeal and vagus
nerves are seen emerging (fig. 2) ; under the ampulla of the horizontal canal the
facial nerve (7) is seen to pass across under the otic process of the suspensorium
(ot.p).

In the angle between the skull-wall and the ear-sac the great Gasserian ganglion (5)
can be well seen, its first branch passing beneath the ascending process of the sus-
pensorium (figs. 1, 3. 5, a.p) ; and in a notch on the back of the sphenethmoid the optic
nerve (2) runs out. The first nerves pass through the trabeculee much further forwards.

The suspensorium still retains its early position and unossified condition. It is a
large multilobate mass. The transverse condyloid scooped, distal end looks forwards ;
from that part the cartilage thickens rapidly, and. spreads into three processes, two
running inwards, and one outwards and backwards. Of the inner processes, the upper,
or ascending process {a.p), is the smaller ; it coalesces with the side of the skull-wall,
near the top.

The lower inner process is the pedicle (fig. 2, in front of 5) ; it is much larger, and its
round end pushes inwards, without coalescing, in front of the ear-sac ; all the branches
of the trigeminal nerve pass over this process. The outer or otic process (fig. 1, ot.p) is
raised into a curved ridge above, and this elevation winds round the front of the auditory
sac, halfway to the skull-wall, taking the same curve inward as the ascending process.

The pterygopalatine bone {p.pg) is somewhat constricted behind the teeth, and ends
behind in a subflabelliform flap, which binds the suspensorium below ; but there is no
pterygoid process of cartilage growing from the suspensorium. Therefore this skull
corresponds, here, to the 1st stage of the Salamander's (PI. XIV.) ; and it also corresponds
with the lower Perennibranchs in form of the vomer in front, which is a long falcate
dentigerous plate, converging towards its fellow.

The premaxillary is an elegant azygous bone, with its dentary margin arcuate, and
having from its top, at the middle, a thick ascending process, which ends in two long-
lathy nasal processes (n.px), that overlap the frontals (f) halfway to the parietals (p).
As in many Urodeles, the sharp notch between these two processes has in it a rounded
tubular cavity, the "median nasal passage," whose counterpart is to be seen in the
Lamprey and Hag-fish ; it is a non-functional, historical space.

The frontals (fig. 1, /) overlie the parietals (p) as much as they are overlain by the
nasal processes of the premaxillaries.

In this embryonic condition the vomers are wide apart behind, and the trabeculae are

180 PEOF. W. K. PAEKEE ON THE MOEPHOLOGY

half-naked between them, and the large basal splint, the parasphenoid (fig. 2, pa.s).
This bone reaches more than halfway to the fore end of the long vomers, and nearly to
the rim of the foramen magnum ; it is narrow, and for the foremost two fifths deeply
split, a small third lobe lying between the two sharp processes. This is a rare condition
of this bone, which is only slightly split or gently emarginate as a rule.

A thin subcutaneous tract of bone binds the suspensorium on the outside. This is the
squamosal (sq) ; at present it is very ichthyic, and is mainly composed of the preopercular
portion. Several other bones appear afterwards ; we shall find them in tbc skull of the
adult specimen.

The inferior arches (PI. XVII. fig. 4-) are now in their most perfect condition, the
three jrill-bearers beincr in full function : thev arc in all essentials like those of
Salamandra maculosa (PI. XIV. fig. 3).

The mandible is lodged in a trough-like articularc (ar) ; a long splenial (sp) runs along
the inner side, and the outer face is invested by the dentary (d).

The hyoid (c.hij) is shaped like a scythe-blade ; it is thickish, especially below. At
that end a short hypohyal (//./>//) is segmented off.

These are the four normal branchial arches, the last of which is functionless. According
to rule, the two first only have a ceratobranchial element (c.br1^2), the rest are merely
epibranchials (e.br'&i). The 1st basal piece (b.br1) receives the 1st and 2nd cerato-
branchials by articulation ; the hypohyals arc attached loosely by ligament ; the 2nd
basibrancliial (b.br) carries nothing, lies on a lower plane than the 1st, and has its
distal end pedate.

The Skidl o/'Notophthalnius viridescens. Second or adult staye.

The adult of the specimen dissected by me measured 3j inches from snout to end of tail.
I had no other stages to examine but these two, the larva and the adult ; but they are
possessed of the true characteristics of an old and a young, or of a low and a high, Uro-
dele. This skull is not unlike that of our native large Newt (Triton cristatus), and
differs considerably from that of the viviparous Salamander (PL XVI.), being longer
and stronger.

Here, indeed, and in some related types, as in my next instance, Cynops (PI. XVIII. ),
we have a very fine piece of skull-building, a bony cranium, forming as safe a structure
as could well be provided for any creature whose safety is founded on " the brittle
strength of bones."

It is an almost perfect osteocraniwu, cartilage persisting merely where cartilage is a
necessity to this kind of organization. The halves of the occipital arch meet above
(PI. XVII. fig. 5) ; but there is a small patch of cartilage on the threshold (fig. 6).

The small crescentic condyles (oc.c) are but little seen above, as they look principally
downwards and backwards ; they are wide apart, having a round notch between them for
an intercalary vertebra.

The ear-masses are large and very strongly ossified ; the large posterior canal (fig. 5,
p.sc) runs along nearly parallel with the occipital rim to meet the anterior arch (a.sc).
The two arches join at a right angle, and the foremost has a crescentic ridge traversing

OF THE SKULL IN THE AMPHIBIA ITKODELA. 181

it, whose convexity is backwards. The ampulla of the horizontal canal (h.sc) is hidden
by this ridge, for its outer end joins the squamosal (sq). That canal lies a in strong roof
(" tegmen ") laterally ; this is enlarged by the squamosal ; and the soft end of the " otic
process " (figs. 6, 7, ot.p) lies lodged between the two bones. The elegant bulging of the
vestibule in the same figures forms a lozenge-shaped convex tract below, bounded on the
inside by a sulcus, and on the outside by the " tegmen." A smaller lozenge exists in
the hind part of the larger ; this is the cartilaginous stapes (st), fastened to the rim of
the 4-sided fenestra ovalis. At the antero-internal angle of the vestibular swelling, just
outside the basitemporal projection of the parasphenoid, the 7th nerve escapes.

The whole side wall, or sphenethmoid (figs. 6, 7, sp.e), is ossified from the prootic
(pro), with which it is ankylosed, to the nasal sac. At its hinder fourth the optic nerve
(2) escapes ; then there is a lesser foramen for some of the orbital nerves, and close in the
axil, between the wall in front and the ear-sac behind, the 5th nerve escapes. Part of
the unossified nasal roof (na) can be seen forming a curtain to the external nostrils
(e.n), and another part behind and below, attached to this, and most probably confluent
with it, is the ethmopalatine (fig. 6, e.pa).

The suspensorium is much less retracted than in the Salamander (Pis. XVI. and
XVII.) ; the pear-shaped condyle is in front of the foramen ovale. It is nearly all
ossified by the quadrate (q) ; the upper part is completely hidden by the pterygoid (pg) ;
but the otic process (ot.p) remains as a tongue of cartilage attached to the grooved
" tea-men."

The pterygoid (pg) has its short part swollen towards the skull ; this is seen to cover
the pedicle, when the skull is examined from below (fig. 6). The outer margin strongly
binds the inner face of the quadrate, and tbe bone from thence becomes a blunt style,
attached by a short ligament to the maxillary (mx).

The dentigerous segment of the larval pterygopalatine bone (figs. 1, 2, p.pg) is com-
pletely confluent with the vomer of the same side (fig. 6, pa, v). The space between
these palatal bars is narrow, but gradually widens backwards.

About a fourth of the bone (less on the left than on the right side) becomes seg-
mented off as a postpalatine bone. I call this a postpalatine because I cannot reconcile
it with the mesopterygoid *. It may, however, be the same bone, for it comes off from
the same part of the palatine as that which gave off the pterygoid. At any rate the
palatine itself, here, only answers to that curved ridge of a bird's palatine that cleaves
, to the parasphenoidal rostrum.

Where these palatal beams widen there they become vomerine ; that widened part,
with the ethmopalatine cartilage, bound the internal nostril (i.n) behind. The bone
narrows again to form the inner rim of the internal nostril, and then grows into a
hatchet-blade, whose back has a round notch. The two notches together form an oval
fenestra.

The outer bones of the fore face have an equal development of hard palate, as we see

* In a generalized Woodpecker (Picwmnus minuius, see Trans. Linn. Soc. ser. 2, vol. i. pi. 5) I found a distinct
postpalatine bone. In that group the mesopterygoid is not fretted off, as in most " Carinatoe."

SECOND SERIES. — ZOOLOGY, VOL. II. 25

182 PROF. W. K. PARKER ON THE MORPHOLOGY

in the two vomers. The premaxillary (px) yields a very large transversely oblong
plate to the palate in front ; the dentary margin is only gently arcuate, and the
ascending nasal process (figs. 5, 7) is a smallish wedge. The maxillary (mx) gives off
a considerable palatine plate, and is of great height in the cheek (fig. 7) ; its long jugal
process is arched downwards, and only reaches two thirds of the distance from the high
part of the bone to the quadrate (q). These bones, as also all the subcutaneous scutes for
this strong skull, arc rough on their outer surface.

The nasals (n) are large, thick, scabrous, multilobate plates ; their strong median
suture is notched before and behind, — before to receive the apex of the nasal process
of the premaxillary, and behind to form, with the frontals (f), a small fontanelle.
This space communicates with the more regular opening between the vomers below;
and this merely functionless cavity is the adult remnant of the "middle nasal passage"
seen in the larva (figs. 1, 2).

The frontals (/) are bevelled at their antero-external angle, to receive the outer nasal
roof-bone, or external ethmoid (e.eth). This is a roughly pentagonal plate; it cor-
responds with the bone which ossifies the prefrontal of an Osseous Eish, but is parosteal
in these types.

The orbito-nasal nerve runs in between this bone and the maxillary in the lacrymal
region (fig. 7). The frontals are imbricated on the parietals (p,f) ; they scarcely exceed
them in length, but are broader. This breadth is gained by the development of a much
better orbital rim than is seen in Salamcmdra (PI. XVI.).

The free outcurved part of the rim reaches to the squamosal ; in front, it is
finished by the ectethmoid ; and below, the jugal process of the maxillary almost com-
pletes the bony circle of the eye. The ruga3 and their interspaces, especially between
the inner and outer part of the frontals, very much resemble what is found in so many
Teleostean Fishes. The parietals (p) really form a wall in this species as well as a roof; and
a strong ridge, which converges in a subarcuate manner towards its fellow behind, makes
a very definite upper boundary to the temporal fossa. The size of the muscles causes the
skull to have a waist at that part ; and the fossa itself is completely surrounded by bone.

The parietals end behind abruptly, the temporal ridges in each angle projecting some-
what ; but each bone is lifted and bevelled a little over the junction of the anterior and
posterior canals (a.sc, p.sc).

The frontal and sagittal sutures run on to the end, and in front is the nasal suture,
over the snout ; these are roughly denticulate.

The side view (fig. 7) shows how well made are the orbital plates of the frontals and
the temporal plates of the parietals.

The hammer-shaped squamosals (sq) have their supratemporal equal to their preoper-
cular region ; the former is roughly ribbed along its outer edge above, and the latter is
carinate, the keel running down into the pointed lower end, and binding on the quadrate (q).

The fiat top of the squamosal (fig. 5) is scabrous ; its supratemporal part is twice as
large as its postfrontal projection. A notch behind, under the teginen, exposes the
otic process.

OP THE SKULL IN THE AMPHIBIA UEODELA. 183

There is another facial bone in this species, the small, graniform, septo-rnaxillary
(fig. 7, s.mx).

The forked, narrow, fore end of the parasphenoid (fig. G, pa.s) is largely invested by
the palato-vomerine plates ; further back it gradually expands to the basitemporal angle,
which is short and rounded. Behind this, as the bone narrows backwards, there is
another like projection, and at the middle the bone ends in a rounded manner. This
bone is subcarinate between the palatines ; the keel is grooved along its middle ; and
in the interauditory region the crest swells into a gentle convexity.

Thus we see that even the huge parasphenoid becomes specialized, and even ornate,
in the higher Caducibranchs.

The mandible is arched from chin to angle, both outwards and upwards (figs. 7 & 8) ;
the two bars form a strong and strongly bent bow. The articular part is highest, and this
height culminates in a sort of coronoid process, formed on the outside by the dentary (d),
and on the inside by the articular (ar). The splenial (fig. 4, sp) is a delicate style, most
pointed in front. There is but little cartilage left, and this is mainly articular ; the
condyle is placed obliquely, it looks upwards and backwards.

The hyobranchial apparatus (fig. 9) has undergone similar changes to that of the
Salamander (Plate XIV. fig. 7) ; but there are some remarkable differences.

The ccratohyal (c.luj) has become rounded and ossified in its upper half; its lower half
is very fiat and is soft. The hypohyal (Jt.hy) has become loosened from the main piece,
and now articulates directly with the soft fore end of the first basibranchial.

Like the ceratohyal, the 1st branchial retains its relative size. The two pieces are
soldered together, and yet each has its own bony shaft (e.br1, c.bv1). There is a little
cartilage at the free end of each segment ; and that of the ceratobranchial articulates with
the soft hind part of the 1st basibranchial. A delicate arcuate rod of cartilage articu-
lates with the same tract, but behind it. This is bowed inwards as much as the ossified
bar is bowed outwards ; it is the second cerato-branchial (c.br) whose upper piece has
been absorbed. This rod is fused with the synchrondrosial cartilage of the one in
front of it.

The two other arches have been absorbed ; and so also has the 2nd basibranchial, all
but the lobes at its distal end (figs. 1 & 9, b.br). These are now thick semioval thyro-
hyals. Behind these nuclei, which bound the glottis (gl), there are two more patches of
cartilage (Lv) ; they are roughly-perforate irregular rudiments of the arytenoid cartilages.

All but the ends of the 1st basibranchial (b.br1) is occupied by a shaft-bone, so
that there are seven of these rods in these metamorphosed arches.

Both the Anura and the Urodela curiously foreshadow the hyoid and thyro-hyals of
the higher types ; but the Frog casts its shadow towards the Mammal, and the Newt
towards the Bird.

Species 3. — The Skull of Cynops pyrogaster (adult).
This Japanese Newt is larger and stouter than the last type ; its skull is very massive,
and greatly resembles that of Notoplithahnus, but is still more dense and strong. This

25*

184 PEOF. W. K. PARKER ON THE MORPHOLOGY

form belongs to the same natural group as the last ; and the morphological differences
between the two are, indeed, interesting, but somewhat trivial.

The occipito-otic region (PL XVIII. figs. 1, 2) is strongly in contrast with the rest of
the skull ; the wrhole skull is a sort of short stout mallet, the handle being all the middle
and front regions, and the head the ear-masses, formed into one by the occipital arch
and its over- and underlying splints.

The arch of the occiput leans forwards, so as to show the floor from above. This floor
is widely emarginate, to receive the odontoid vertebra (not process), which is an exact
miniature of the odontoid process of the Ox. The exoccipital bones are very closely
applied both above and below ; the wavy facets (oc.c) are subpedunculate and look down-
wards and inwards. Above, the huge canals of the ear (a.sc, p.sc) elbow inwards to the
rim of the arch ; they leave a hollow outside, which swells again with the horizontal
canal (h.sc) ; this is overlapped by the squamosal (sq). The anterior region of the
auditory sacs and occipital arch is covered by the parietals (p) ; below (fig. 2), the pedicle
of the suspensorium, invested by the pterygoid, hides the ear-mass in front.

The vestibule, behind the foramen for the 7th nerve, is wrcll defined, but does not
bulge so much as in the last kind (PI. XVII.). The fenestra ovalis, in its hinder margin,
looks outwards and backwards. The oval stapes (st), with its nucleus of bone in the
centre, resembles, on a large scale, the blood-corpuscle of a Newt.

The tegmen is irregularly scooped, and is enlarged by the squamosal (fig. 2, sq). The
whole mass is very wide, directly transverse behind, Avith the middle region having the
condyles projecting on each side of the median emargination. The pterygoids, suspen-
sorio , and squamosals, together, grow like trifurcate antlers, from the broad, transversely
oblong, occipito-otic mass.

From the auditory capsules, onwards, the skull-walls are ossified, as " sphenethmoids,"
nearly up to the internal nostril (fig. 2, i) ; the rest remains soft (figs. 2 & 3, sp.e). The
external nostril is well rimmed by the olfactory cartilage (na, e.n). The internal nostril
is margined behind by the ethmo-palatine cartilage ie.pa), which coalesces with the
nasal wall.

The optic nerve (2) passes out at the hinder third of the sphenethmoid ; the 5th (5)
betwreen that bone and the prootic {pro, sp.e), and the orbitonasal pierces the ecteth-
moidal plate (e.eth, 5).

When the skull is seen from its fore end (fig. 6) the large fenestra for the olfactory
crura (1) are seen ; the 9th and 10th nerves (fig. 2. 9, 10) emerge close on the outside
of the occipital condyles (oc.c) ; the first of these burrows in the capsule, but the
vagus escapes in the chink between it and the exoccipital, so that a small bar of bone
can generally be seen dividing the foramen into two.

As in the last species, the suspensorium looks forwards as well as outwards, not gain-
ing that backward position seen in the Salamander (PL XVI.) ; it is well ossified (q), and
has a cylindroidal condyle. The otic process {ot.p) is triangular, soft, and free.

The pterygoid cartilage {e.pg) has been ossified by the pterygoid bone, especially at its
origin, and thus a separate epipterygoid has been formed. The strong pterygoid bone

OF THE SKULL IN TILE AMPHIBIA UEODELA. 185

strongly binds the suspensorium and its pedicle, and then runs directly forwards as a
round-pointed tongue, which does not nearly meet the maxillary, that hone being short.

The palatines are like those of the last kind ; they are long flaps of bone, running back
from their union with the vomers, and are serrated with small teeth along their inner
edge ; the right bar has a separate postpalatine detached from, and one third the length
of, the main piece*.

The left bar reaches further back than the other, even with its additional piece. The
fore palate is not so extensive as in Notophthalmm, the palatine plate of the vomers
being less ; each bone is rimmed round the oval middle nasal space. The premaxillary
and maxillaries (px, mx) are quite similar to the last ; but the jugal process of the max-
illary is much shorter and the bones are stouter.

The nasal process of the premaxillary is grooved and pitted in front ; the building
together of the fore face and nasal openings is shown in a front view (fig. 6, n.px, na, e.n).
On the right side two small irregular septo-maxillaries (s.mx) are seen resting on the
maxillary ; on the left side there is a small grain resting on the alinasal cartilage.

The ectethmoid (e.eth) is a very strong wedge, finishing the antero-superior region of
the orbit, and pierced for the orbito-nasal nerve (fig. 3. 5).

Each nasal (u) has a crescentic notch in front and an oval lobe behind. Their rough
common suture shows two lozenge-shaped gaps that communicate with the internasal
recess.

The frontals (/) are strong, and give off a thick scabrous postorbital process from their
orbital rim and plate ; the squamosal (sq) forms a squamous suture with this process.

The parietals (p) are about equal in size to the frontals, and each bone has a strong
transverse ridge behind the temporal fossa, which is triangular ; it is scooped behind
this ridge. The frontal, coronal, and sagittal sutures are very irregular and coarsely
toothed.

As in the last, the squamosal (sq) is a large T-shaped bone, carinate, both along its
outer edge above, and down its preopercular bar in front and below. The paraspheuoid
(pa.s) gradually widens to the interauditory region, and then suddenly narrows again to
its transverse end close in front of the odontoid notch ; it gently bulges along the basi-
cranial floor.

The mandible (figs. 1 & 5) is a very stout bar, which is arched upwards as well as
outwards. The dentary (d) nearly reaches the condyle outside, and the deep trough-
shaped articulare (ar) helps it to form the coronoid process. The splenial (sp) is a long
f-shapcd splint, following the curve of the dentigerous ridge of the dentary ; it is pointed
at both ends ; the condyle is cylindroidal.

The hyobranchial apparatus closely resembles that of Notophthalmm, but has some
curious differences. Every part is stouter; the hypohyals (h.hi/) look forwards; the
1st epi- and ceratobranchials (e.br1, c.br1) are not fused, either with each other or
with the 2nd cerato-branchial (c.br) ; and this bar is also ossified. The thyrohyals,
formed from the 2nd basibranchial (b.br-), are better formed than in the last; the

* The suture is not drawn in fie. 2.

186 PEOF. W. K. PAEKEE ON THE MOEPHOLOGT

laryngeal patches (Ix) are quite similar. Altogether the skulls of Cynops and of Notoph-
thalmus are only unlike in non-essentials.

A considerable number of the Caducibranchs agree with each other in all that is
essential in their skulls ; but others show remarkable modifications, as I shall soon
show.

Species 4. — Skull of Taricha torosa (adult). North America.

This skull is only two thirds the size of the last ; it is flatter and feebler, but agrees in
all essentials with the stronger specimen. The cartilaginous basi- and supraoccipital
regions (PL XIX. figs. 4, 5, s.o,f.m) are wider, in conformity with the less intense ossi-
fication of the skull generally.

The occipital condyles (oc.c) are distinctly pedunculated, as a correlate of the deep
odontoid notch ; they look a little inwards and downwards, but most backwards.

The auditory masses are one continuous bone on each side with the occipital moiety ;
they look forwards and outwards, and the general outline behind is crescentic, not
straight, as in Cynops. The semicircular canals are strongly marked on the huge ovoidal
ear-masses (fig. 4) ; below, the fenestra ovalis is largely surrounded by unossified carti-
lage ; and the oval stapes (st) is also soft. Above, there is a fissure in the prootic tilled
by cartilage ; this is continuous with the facet to which the pedicle is attached, as in
the adult Frog. Here, outside the hole for the facial nerve (7), the prootic itself forms
a sheath to the raised facet for the pedicle (pd), which is hidden by the rim of the
pterygoid (pg).

Prom the middle of the optic passage (2) to the foramen ovale (5) there is a carti-
laginous (alisphenoidal) tract, and then the sphenethmoid (sp.e) runs nearly to the curved
ethmopalatine (e.pa).

The nasal capsules (na) are entirely unossified ; the external nares (e.n) are small and
round, the internal nostrils arc large (fig. 5, i.n). The internasal lamina can be seen both
above and below (figs. 4 & 5), forming the hinder boundary of the internasal recess (m.n.p).

There is a strong quadrate bone (q) in the suspensorium. The condyle looks forward,
as in the last two kinds. The fore part of the suspensorium. gives off a long pterygoid
process, which is longer than the pterygoid bone ; from its base upwards the suspensorium
is unossified, its pedicle being closely locked within the pterygoid. This species has the
condyle of the pedicle most like that of a Prog. Prom the rim of this joint the ptery-
goid bone runs inside the quadrate, wedge-like, and then forwards, supporting the
counterpart cartilage ; the latter touches the zygomatic process of the maxillary (mx),
but the apex of the bone docs not. The falcate palatines (fig. 5, pa) are imbricated
beneath the vomers (v), and the convex edge is not a sharp blade, but a saw, the teeth
being uniserial and inturned (on the right side of fig. 5 the palatine is represented as
absent). The concave back of this little saw corresponds with the outer edge of the
parasphenoid (pa.s).

The hard palate is not so well developed as in the last two kinds. The vomers (fig. 5,
v) are rough miniatures of a pruning-hook, the semicircular notch margining the inner
nostril (i.n). The convex back of the hook is articulated by suture to the palatine plates

OF THE SKULL IN THE AMPHIBIA UKODELA. 187

of the premaxillary and of the corresponding' maxillary (px, mx). The "helve" of
the blade is the palatine, and each blade is wrought into sutural teeth to unite, by squa-
mous suture, with the underlying fore end of the palatine, which carries, now, all the
true teeth; for the vomer has become edentulous.

We have thus, in this species, a most instructive analysis given us of the skull of the
two last kinds (compare PL XVII. fig. 6, PI. XVIII. fig. 2, and PL XIX. fig. 5). More-
over, this skull, which stands related to the last two, like a young individual, is also
intermediate between them on one hand and Desmognathm and Spelerpes on the other
(PL XXL).

In AmUystoma (Phil. Trans. 1877, pi. 21. fig. 6) the palatine, as in Polypterus and
the Batrachia, is a transverse bone, which, with the ethmo-palatine, its cartilaginous
consort, represents the ornithic " os uncinatum." In Spelerpes and Desmognathm
(PL XXI.) I shall describe a palatine which is both transversely and basally developed,
as in those Batrachia which have a broad inner part (e. g. Bnfo agua).

In Taricha (PL XIX. fig. 5) the palatine {pa) has no transverse part, but it is longi-
tudinal or basal, altogether ; yet it does not run much more than half as far backwards
as in the last two kinds.

The median palatine region of the premaxillary (px) is a grooved emarginate wedge.
The nasal process (fig. 4) is a very curious structure ; there are two spatulate free pro-
cesses, but these are confluent from the middle forwards ; and this part is largely sulcate,
the valley being surrounded by a ridge, which is round in front, and which °at the sides
passes into the free lobes.

The uniting portion ends in a rounded notch, which is converted into a large oval
hole by the internasal cartilage behind. The same cartilage is seen to bound it below
(fig. 5, m.n.p), between the hooked vomers.

In this careful finish of a useless middle cavity or passage,. Taricha approaches
Spelerpes and Desmognathm (PL XXI.) ; those surroundings are not less perfect in the
last two kinds (Pis. XVII. & XVIII.).

The maxillary (mx) in this flatter skull has a lower facial plate ; its palatine portion
also (fig. 5) is less extensive ; but its jugal process is unusually long, passing some
distance behind the pterygoid (pg). The septo-maxiUary is not apparent.

The nasals (») are more rounded and thinner; the orbital process of the ectethmoid
(e.eth) is a sharper spur, but forms part of a very similar superorbital ridge ; that part
of the frontal (/) articulates over the temporal fossa with the squamosal (sq).

The frontals do not quite cover the internasal cartilages, and on each side they are
dinted by the nasal processes of the premaxillary, and overlapped by the nasals and
ectethmouls. In their turn they are imbricated upon the parietals (p), which run on
between, as well as beneath them.

Together the parietals (p) form a roughly pentagonal plate, divided irregularly by the
sagittal suture.

Each bone is imbricated upon the prootic over the anterior canal (a.sc), and the two
run straight across the occipital arch, a little in front of the shelving upper ed<*e of the
foramen magnum.

188 PEOF. W. K. PAEKEE ON THE MOEPHOLOGT

The parasplienoid (pa.s) is essentially like that of Cynops. The postorbital part of the
large squamosal is almost as large as the part which lies over the two anterior ampulla} ;
the fore part curves inwards to join the postorhital process of the frontal. A carinate
preopercular process runs down outside the suspensorium, binding the quadrate (PL XIX.
fig. 5, sq, q).

The mandible (figs. 6 & 7) is feebler, lower in the corouoid region, and has a smaller
splcnial bone (sp) ; the condyle is a large oval tract of cartilage. Each ramus, as
usual, is tied to its fellow at the chin by a strong interosseous ligament.

The hyobranchial apparatus (PI. XIX. fig. 8) might have belonged to a half-grown
Cynops ; the bony centres are shorter, only half as much developed, and there is no
bone in either the first basibranchial (b.br1) or on the second cerato-branchial (c.br).
Another arrested state is shown in this, that the Avhole end of the second basibranchial
(b.br") remains as a transverse thyrohyal, and is not broken up into two segments, as in
the last two kinds.

Thus this species, conforming on the whole very accurately to the highest Caduci-
branchiate type, is, nevertheless, somewhat arrested in development as compared with
some other kinds.

Species 5. — The Skull of Onychodactylus (? species). Japan. Total length of animal

2£ inches.

The study of tins type wrought much surprise in me : to all appearance it seems to be
a genuine Caducibranch, yet in development it has not advanced much further than in the
ripe young of Salamandra maculosa (PI. XV.); and its position in the " Order" would
appear to be among the Cryptobranchs, or those whose branchiae arc almost absorbed.
When I come to describe the skulls of the Menopome and Siren, it will be seen that in
many respects they are quite as much metamorphosed in their skull as this species.

The sharp toe-ends of this kind suggest the presence of nails ; but I could find no
special development of the epidermis on the upper surface; yet I carefully examined
the toes at Prof. Mivart's request, to whom I owe this and many more Amphibian
treasures.

The cranial territories are very equally composed of cartilage and bone ; and the
secondary bony plates correspond to those of the larvae of Caducibranchiate skulls — larvae
that have their gills full-sized or in full play.

The existence of so small an Amphibian as this lingering on the headland of the lowest
tribes is very interesting, more especially as it coexists in Japan, if not with the lowest,
yet with the largest of these waifs of a subextinct group. I refer to Sieboldia. In
South Europe we have the loioest, namely Proteus anrjuinus ; but this creature is an
"albino " prisoner in the waters under the earth. Yet, with the exception of this kind
and of the Gigantic Salamander of Japan, I am not acquainted with any other species
of the lowest group outside North America.

There, still flourish the Menobranchus (which is next above the Proteus), the Meno- \
pome, the Siren, and the Amphiuma ; the small true Caducibranchs (they are all of
small size) have a wide geographical distribution.

OF THE SKULL IN THE AMPHIBIA UEODELA. 189

In this chondrosteous skull the exoccipital (PI. XIX. figs. 1 & 2, e.o.) runs over the
epiotic and opisthotic regions, reaching the fenestra ovalis helow (fig. 2, fs.o), and
covering the arch of the posterior canal (fig. 1, p.sc.) ahove.

An ohlong haud of cartilage finishes the roof and the floor of the occiput (s.o, b.o)
and these tracts pass further forwards into right and left tracts of cartilage that inter-
vene between the exoccipitals and the prootics (pro). These latter bones are less deve-
loped above than below ; above, they skirt the fore margin of the auditory mass, and
then leave the arch of the horizontal, and the ampulla of the posterior, canal unossified
(fig. 1, h.8C,p.sc). Below (fig. 2), the bony tract has grown to the outer margin, but
the vestibule is soft ; thus the fenestra ovalis has half its rim unossified. The 9th and
10th nerves (9, 10) are well enclosed in bone, which, as usual, runs between them.

The position of the fenestra ovalis (the stapes was lost in dissection) is postero-
external. The sphenethmoids (sp.e) barely harden half the intcrorbital walls ; they reach
neither to the nasal region in front nor to the optic foramen behind ; thus the whole
alisphenoidal region and the hind part of the orbitosphenoidal is unossified (fig. 2) ;
so, also, is all the nasal region, which is floored by very large leafy trabecular cornua
(c.tr), and bounded behind by the ethmo-palatine cartilages (e.pa).

The condyles of the suspensorium look forward and reach nearly to a line drawn
across the optic nerves (2) at their exit; there is a quadrate bone (q), which, however,
does not ossify half the cartilage. The otic process (ot.p) is not hidden by the pterygoid
bone, but the pedicle is. This species agrees with the Menopome in having that bone
(fig. 2, p.pg) divided from the palatine. It is a curious sort of small hatchet, with a
rounded blade that binds upon the inner face of the suspensorium, and a haft which is
bent outwards, and ends in a sharp point in front, which has no teeth on it. The large
vomers (p) are perfectly larval, and are confluent with the small palatines ; they meet in
front, and then form an elegant arch, whose anterior margin carries teeth. The inner
end of each bone is rounded, and the outer sharp, the sharp end overlapping the pterygo-
palatine. Just in front of the middle of each plate there is a rounded inner lobe.

This arch is behind the large flabelliform cornua trabecule. In front of those lobes
there is a larger dentigerous arch, very elegant and forming nearly a semicircle. All
except the extremities of this bone is formed by very large premaxillaries ( px), the ends
by small maxillaries (mx). As in Proteus, Menobranchus, the Siren, Menopome,
Axolotl, and Spotted Salamander (Pis. XIV.-XVL), there are two premaxillaries ; but
most of the Caducibranchs have but one, as in the last two instances.

The nasal process of each bone is slender, and the palatal edge of each bone is
■narrow ; the maxillary (mx) is a small style attached to each outer end. The nasals
(fig. 1, n) are oblique lozenge-shaped bones, pointed in front, and to a less degree
behind ; the nasal processes of the premaxillaries rest on them, they on the frontals (/),
and the frontals on the parietals (p), by a normal regular imbrication.

A conchoidal ectethmoid (e.eth) flanks each nasal, and also partly the outer edge of
the frontal. The interorbital region of the skull being very long, the frontals and
parietals are very long and narrow ; the latter run far forwards as a wall-plate to the
skull, and are largely overlapped by the frontals.

SECOND SERIES. — ZOOLOGY, VOL. II. 26

190 PEOF. W. K. PARKER ON THE MORPHOLOGY

The supratemporal region of each parietal is hilobate ; it lies over the anterior canal
(a.sc) ; behind this expanded part each bone slowly lessens towards its fellow at the
mid line ; they leave much of the supraoccipital region naked.

The squamosals (fig. 2, sq) are mere splints binding the suspensorium, and running
along the side of, and somewhat above, the ear-mass.

On the floor of the skull we see a very long, delicate, inferiorly bulging parasphenoid
(fig. 2, pa.s) ; in the space between the fore part of the two vomers it is very splintery ;
it gradually narrows to the foramina ovalia, and then spreads into a large ox-face of
bone. This basitemporal part has a pair of large rounded lobes, with a lesser pair both
in front and behind these : the end is transverse and somewhat emarginate ; it reaches
much further back than the parietals.

The mandibular arch (fig. 3) is a delicate and very elegant bow, much unlike the
same arch in the stout Newt just described, and very larval in character. The Meckelian
core of cartilage (mk) is unaltered by its investments ; it reaches the chin inside the
dentary (d), and its thick hinder part lies in the articularc (ar) as in a trough. Tbat
bone is scarcely more than three fifths the length of the mandible, and the splcnial (sp)
is very small and lies on the fore end of the articulare. The condyle is oval and large,
as in larval Caducibranchs ; it rolls loosely on the slightly scooped condyle of the sus-
pensorium.

As in the Menopome and Siren, the hyobranchial apparatus has undergone but little
change ; there has been no absorption of the hinder arches, and their denticulations
remain : these things all show that this type is not a true Caducibranch.

The hyoid and 1st branchial are much alike, but the latter is the longer and slen-
derer ; they both remain soft. Their subdivision differs ; for the ccratohyal (c.hy) cuts
off a hypohyal (h.hy), here unusually large ; but these two together correspond morpho-
logically with the ccratobranchial ; for the epihyal is suppressed. The thick 1st cerato-
branchial (e.br) is attached to the middle of the 1st basibranchial, a long oval cartilage
(b.br1), and the 2nd ceratobranchial (e.br2) to its end. This last bar is as long as
the one in front of it, but is very much slenderer ; it also has acquired an imperfect
bony shaft. Its epibranchial piece (e.br2) carries the 3rd and 4th arches, which are
composed, as usual, of merely the epibranchial segment (e.br'i & 4). The 2nd basi-
branchial (b.br2) has begun the usual metamorphosis ; its fore half has become mem-
branous, and the hind half is ossified, all but the outer lobes, the only parts that are left
in Notophthalmus and Cynops (Pis. XVII., XVIII.) . The presence of three bony
centres, and the absorption of the hindermost median bar in front, are the most marked
evidence of metamorphoses of these bars.

Considering our great poverty as to existing Perennibranchiate and Cryptobranchiate
Urodeles, this curious sharp-toed Japanese Salamander turns out to be a form of no
little importance. It is a very useful stepping-stone from the lower to the higher kinds
of its own " Order."

OF THE SKULL IN THE AMPHIBIA UEODELA. 191

Species 6 Sr 7. — The Skulls (larval and adult) of Spelerpes rubra, and larval skulls of

Spelerpes salmonea.

These North- American Caducibranchs measured as follows : — The adult S. rubra 4^
inches long, and the young (with gills) 1 inch 10 lines ; the larval specimens of S. salmonea
measured 2 inches 1 line and 3f inches. The larva3 of this genus reminded Prof. Mivart
(' On the Common Prog,' p. 48) of Menobranchus ; they struck me as being like Proteus.
We were both right ; for they strongly resemble and very accurately represent both those
Perennibranchs. The youngest of these larvse (S. rubra) was of considerable length,
namely If inch, and yet the structure of its skull is extremely simple, and differs in
nothing essential from that of the adult Menobranchus lateralis (compare PI. XX.
figs. 1 & 2 with that type : see Huxley, P. Z. S. 1874, pis. xxix. & xxx.).

This larval skull is curiously illustrative of the extreme variability of the time in
which ossification is set up, and harmonizes with what I find in the Batrachia, namely,
that of the most generalized kinds, some skulls are most feebly ossified, and others
are a mass of dense bone, with scarcely a suture left.

Here there is no distinction even between the prootic and exoccipital ; all the occipito-
otic region on each side is one undistinguishable tract of bone (PI. XX. figs. 1 & 2, pro,
e.o) ; and these bony territories have met in the supraoccipital region, although they
have a broad synchondrosis below. There is a character here which strikes the eye at
once, namely, that the foramen magnum is nearly as wide as the average width of the
cranial cavity. This is surely a low character.

The condyles (pec) look obliquely upwards, like ordinary " pre-zygapophyses," and
between them there is a very wide crescentic notch for the " odontoid vertebra." There
is a rudiment of a bony roof to the skull running forwards from the supraocciptal region.
This bony tract first grows inwards as a lobe, and then stops short over the Gasserian
ganglion ; thence the skull-wall is unossified, and only behind does it lean over the
brain.

The whole configuration of the semicircular canals (fig. 1, a.sc, p.sc) is stamped upon
the bony periotic. These structures are very large.

It will be seen that above, in front of the two ampullae, there is a tract of cartilage ;
this is an upgrowth from the basal plate (trabeculo-parachordal), and is normally
related to the pier of the mandibular arch, the " suspensorium." On the lower face
of the ear-capsule there is some cartilage left ; this is a trilobate tract beneath the
" sacculus ;" outside this tract is the large fenestra ovalis (fs.o), with its correspondingly
large stapes (st).

A considerable blood-vessel is seen entering into (? emerging from) the membranous
labyrinth at the inner edge of the regularly ovoidal stapes, whose narrow end looks
forwards and outwards.

There is a short canal for the facial nerve (fig. 2. 7), but it soon escapes from the
capsule, and runs outwards and forwards beneath the apex of the otic process (ot.p), some
distance in front of the stapes.

The 9th and 10th nerves (9, 10) are seen escaping from the skull through their twin-
passage, between the condyles and the ampulla of the posterior canal (p.sc).

26*

192 PROF. W. K. PARKER ON THE MORPHOLOGY

Prom the front of the ear-capsules to the frontal wall there is nothing hut a pair of
cartilages, the trabeculse ; these hulge externally, are concave within, are of moderate
height, that is, they are as high as this flat skull is within, reaching ahove to the frontals
and parietals [f,p), and helow to the parasphenoid (pa.s).

As they approach the nasal sacs they become flat and depressed ; having lost their wall
character, they became part of thejloor only. Each plate of cartilage is bowed outward
twice — first in the alisphenoidal region, and again in the ethmoidal ; but on the whole
they gently approximate, but are a good distance apart at their end. Then they are
first rounded and then each is Battened out into a small flabelliform " cornu " (c.tr).

Close behind these cornua the trabecule are undergoing conjugation, each bar
sending inwards a small internasal lamina (i.n.c). Between the body of the premaxillary
and these processes there is a large internasal cavity (m.n.p).

In similar larva? of Salamandra maculosa (PI. XIV.) the notochordal region was well
floored with cartilage, the large expanded ends of the trabecular having coalesced with
equally large parachordal tracts. Even in the adult of that kind there was still the
trabecular bridge, then a fontanelle, and then the true parachordal tract. Here, however,
so early in life, the parachordal cortilages only form a conjugating belt over the end of
the parasphenoid (pa.s). Thus it is evident that all the inturned parachordal part of
the trabecular and of the fore half of the hinder cartilages, or true parachordals, have
been entirely absorbed, leaving the vicarious parasphenoid to make the floor.

I have watched this process in larva? of Triton cinstatus and mJSTotophlhalmus (Pl.XVIL).
In this latter kind two larvae of precisely the same size differed greatly ; in one (figs. I & 2)
the trabecular ends were reduced to a small lobe, and in the other (fig. 3) they were still
of good size, but they had become quite free from the hinder cartilage.

In both Proteus and Menohranchus the adult skull is much simpler than the embryonic.
My dissection of the latter kind shows more basal cartilage than the one dissected by
Professor Huxley (loc. cit.). In small young specimens I am confidently expecting to
find a much fuller development of the basal cartilage than is to be seen in the adult *.

At present, as in the adult Proteus, there is, as yet, no olfactory cartilage ; the sacs
lie on and outside the trabecular. Behind the internal nostrils (i.u) there is on each
side a free cartilage having a totally different morphological meaning to the nasal
sacs, namely the ethmopalatine (e.pa) ; it is a small semioval free bud, attached to the
trabecula at its postnasal bend, and is the orbito-nasal landmark.

The postorbital region is marked off by the large suspensorium, whose condition
suggests considerable ripeness in this embryo, much in contrast with the slow meta-
morphosis of the fore part of the head.

The condyle is opposite the optic foramen ; the cartilage next to it is ossified as a
quadrate (q), but from the middle all is soft. The true swinging-point is bilobate,

* I am not aware that any brain-bearing vertebrate shows a more simple cranium in its early condition than the
larvae of the Urodeles ; but they possess every thing that is wanted to make up a " ehondocranium " in any type.
Birds and mammals have nothing but the two " parachordal " cartilages, passing into the two " trabecular cranii,"
and the two " intertrabecular " tracts ; and even these are, I suspect, merely differentiated, the parachordals from
the trabecular, by the latter getting the start in chondrification.

OF THE SKULL IN THE AMPHIBIA ITEODELA. 193

vertically ; the upper lobe is the ascending process (fig. 1, a.p) ; it has coalesced with
the alisphenoidal crest near the top.

Tbe lower lobe, or pedicle (fig. 2, pd), is much less distinct than the upper, and seems
like a swelling on the under face of that cartilage ; the orbitomasal and Vidian nerves
pass between this swelling and the skull-wall under the ascending process. The outer
or otic process (ot.p) applies itself, like a snail's foot, to the antero-external face of the
ear-capsule ; it is rounded below (fig. 2), but raised into a distinct ridge above (fig. 1),
which reaches to the anterior ampulla (a.sc). In this stage the otic process is quite
like what is seen in newly metamorphosed Batrachia, notably in Pseudis paradoxa.

The suspensorium is confluent with the skull only by its ascending process ; the other
process, right and left, plays freely on the auditory capsule, and is set in a bed of
delicate connective tissue. The fore margin of tbe suspensorium shows no pterygoid
process, although the ethmo-palatine cartilage (e.jia) is quite apparent.

The lozenge-shaped dentigerous palatine bone has already sent backwards a large, flat,
edentulous pterygoid process, so that we have the widely distributed ptery go-palatine
bone (p.pg) *.

The whole bone reaches from the quadrate to the vomer (fig. 2), and the two bones
with the two vomers (v) mark out half a long ellipse. The vomers are falcate, and nearly
meet in the middle by their broad hooked end ; they are dentigerous, and are about half
as wide as the pterygo-palatmes. In front of the vomers we have the median nasal
passage (m.n.p) in the middle, and the trabecular cornua (c.tr) at the sides.

In front of these there is the dentary margin of the premaxillary (which is not double
as in Proteus and Ilenobranchus) ; above (fig. 1) its nasal processes (n.px) are large and

* As the parasphenoid is the great primordial basal bony plate, so is this the first of the palatals, as we see in
the Dipnoi, which have so few bony lamina} applied to the solid chondrocranium, a step beyond the Ohinneroids and
ordinary Selachians. Higher up it is curious to see the persistency of the pterygo-palatine. In the generalized
Hatteria (see Giinther, Phil. Trans, part ii. 1807, plate i. fig. 2) the huge pterygoid has a large flat anterior
part, which reaches and articulates with the corresponding vomer. In Anguis fragilis this fore part of the
pterygoid boue is separate as a " mesopterygoid." The true longitudinal palatine of the Hatteria lies outside the
vomer and pterygoid.

In Birds tho pterygoid grows forward to the vomer (generally composed of a right and left bony centre) ; but, as
in Anguis fragilis, there is afterwards a separate piece cut off as a mesopterygoid. In all the Carinata), except the
Woodpeckers and the Galline and Anserine groups, this takes place ; and, oddly, the new bone soon coalesces
with the inner edge of the great palatine, which is longitudinally placed, lies outside the pterygo-vomerine
6eries, and reaches from the pterygoid to the front of the nasal labyrinth. I have shown that Picumnus minutus
has a postpalatine, like Cipwjis and Notophihalmus ; and the Passerinoe have an additional cartilage besides
the ethmo-vomerine and pterygoid (the pterygoid of a bird is formed in nascent cartilage and is cartilaginous behind) ;
and this transpalatiue occurs in Siredon. In the Bird it ossifies separately, and then coalesces with the outer part of
the great palatine, as the mesopterygoid does with the inner edge.

Further, the ethmo-palatine of the Frog, when the transverse bony plate has grafted itself on the cartilage,
corresponds with the os uncinatum of the Bird (C'on/tJiai.v, Trogon, Psittaeus) ; and this anterior transverse bone is fore-
shadowed in those Urodeles that have a transverse bony palatine running under their ethmo-palatine cartilage, as in
Amblystoma, Spelerpes, and Desmognathus. Stepping, then, across the wide space between the Dipnoi, the low Urodeles.
and the larva? of the higher kinds to the Bird, we have one safe determination to stand upon, namely, the relation of
the primordial pterygo-palatine to the corresponding vomer. The other parts that come in, in tho metamorphosis of
the Urodeles, or in the ascent up the tribes, are new things — sjiecializations.

194 PROF. W. K. PARKER ON THE MORPHOLOGY

lanceolate. The posterior nares (i.n) are in their usual relation to the vomers; but there
are no maxillaries outside, nor nasals and ectethmoids above.

The two pairs of roof-boues (f,p), with the nasal processes of the premaxillarieSj are
adapted to each other by wedge-shaped imbrications ; the frontals reach the internasal
cartilage, and the parietals overlap the occipital belt slightly (fig. 1).

The parasphenoid (fig. 2, pa.s) is oblong, with rounded corners, and is further irregular
by increasing somewhat from before backwards : the vomers underlie it a little in front;
it reaches nearly to the foramen magnum behind.

The squamosals (sq) are nearly equal in their supratemporal and preopercular regions.
It will be seen that I have figured one and part of another vertebra. The first of these
is imperfect, having no arch ; in the other the arch has been cut away. This second
joint carries the skull on its long pedunculate " pre-zygapophyses," and between these
the curious elongated centrum of the intercalary (odontoid) vertebra is fixed. It is
ankylosed to the second centrum behind, and in front carries a bilobate mass, having a
form somewhat like dumb-bells.

As this intercalary vertebra is inconstant, and as it may represent part of the occiput
in the Sauropsida and Mammalia, for their hypoglossal is a cranial nerve, but is
spinal in these low types, I shall describe this segment the more carefully.

In these types the notochord shortens rapidly (figs. 4 & 6, tic), whilst in Seironota the
cephalic part is large, and partly segmented into two imperfect transient cranial centra.
But if the cephalostyle is absorbed in Spelerpes, the notochord makes up for it in the
intercalary segment. In the large semicircular notch between (not behind) the occipital
condyles a third pair of "parachordal" cartilages appear (besides the ordinary pair
and the parachordal region of the trabecular). These nuclei of cartilage become ossified,
and then coalesce by a bridge of bone; hence their dumb-bell shape. Meantime the long
tract of notochord has become ensheathed with bony substance, and its apex embraces
the bony stolon between the lateral lobes, and coalesces with it, its own end being
emarginate, and more developed below than above the transverse plate.

Here we have the three normal elements of a vertebral centrum ; and this curious
structure is in front of the atlantean segment, a perfect and large vertebra, made to
carry the skull. Thus the human odontoid process, which is merely the ossified noto-
chordal core of the " atlas," enucleated from the symmetrical sides, and ankylosed to the
" axis," is here represented and foreshadowed by a distinct mcta-somatomic joint, with
middle and side elements combined.

The inferior arches scarcely yield in interest to the skull itself, and show that these
types are certainly related to the lowest sort of Perennibranchs.

The cartilaginous pith of each mandibular ramus is very solid, especially towards the
articular end ; that part rests in a delicate bony trough, the articulare (PI. XIX. fig. 9,
ar), and the outside is invested by the bony element (d). Where these bones meet on the
inside, just in front of the highest or coronoid part, there is a small dentigerous plate, the
splenial (sp). Here the splenial is very small ; in Proteus it also is so, and that
type has no hypohyal. Here it is only indicated by a groove crossing the inner face
of the bar below {c.hy, h.hij). The branchials are normal as to form, but abnormal as to

OF THE SKULL IN THE AMPHIBIA UKODELA. 195

number. Proteus and Menobranclms have only three ; the 4th epibranchial is also
wanting here. Thus in several points these larva? approach the lowest " Peren-
nibranchs."

Notwithstanding the large size of this larva, the flat 2nd basibranchial (h.br2) is
very imperfectly differentiated from the 1st (b.br1) ; as to form, these parts are quite
normal.

Skull of larva of Spelerpes salmonoa, 2 inches 1 line m length.

The metamorphosis of this type of skull has been as profitably followed in the larvae of
two species as if I had possessed all of them of the same kind. The head of the one
just described is broader, and more like that of Me nol ranch us ; in this species (PI. XX.
figs. 3-6) we have a nearer approach to the sharp eel-like head of Proteus. Therefore
also, as these are more advanced than the last, we have to deal with characters that are
due to metamorphism, and also with those that are due to specific difference.

The notch between the occipital condyles (figs. 3, 4, oc.c) is shallower ; for the 1st or
odontoid and 2nd or atlantal vertebrae are much shorter than in the last species.

Although more advanced, the supraoccipital region (so) is not yet ossified ; the
basioccipital is more so, and Ave see the very small remains of the cephalic notochord
(nc), so unlike what is to be observed in many Urodelous larva? at this stage.

The occipital roof is of larger extent, and the auditory capsules are directed more
outwards in front, carrying with them the suspensoria, which thus give the larva? a
wider gape.

The amount of cartilage seen on the front face of the capsule from above (fig. 3) is the
same as in the last ; but the vestibular shell has a more even and a narrower rim of
cartilage round the fenestra ovalis, all but its outer side (fig. 4). It will be seen in
these two embryos (figs. 3, 4, 5, & 6) that the " tegmen tympani " is more tilted upwards,
so that the fenestra and its lid, the stapes (st), are brought very close to the outer edge
(see also the difference in PI. XX. figs. 9, 10).

Ossification is spreading from the centre to the circumference of this plate, a neat
margin of cartilage still remaining.

This oval bony centre has two curious enlargements on its surface ; one is a boss on
its hinder half for the insertion of the muscular fibres that roughly represent the
stapedius muscle, whilst the other, on the fore half, is a neat round handle to this
vestibular lid (figs. 4 & 10). As the plate itself has a rim of cartilage, so this rod has a
core at its distal end ; hence it is evident that a distinct rod of cartilage has been
ensheathed by a hollow process from the stapedial bone, just as the prootic ensheaths
the trabecular in these larva? (see especially fig. 6) and in the adult Proteus.

We shall see that this cartilage may have its own osseous centre (see PL XVIII. fig. 9,
in Desitioc/nathus, on the left side only); that is, however, when it does not grow
so close to the stapes. Now here, it would seem, we have at last found a veritable medio-
stapedial. Proteus has an unmetamorphosed hyomandibular ; and what more natural
than that these proteiform larva? should repeat that segment, and use it as a " colu-
mella"?

196 PROF. W. K. PARKER ON THE MORPHOLOGY

In the adult (PL XXI. figs. 2 & 3, st) the stapes lies close to the edge of the tegrnen
and its squamosal tilings, and the facial nerve (7) passes close beneath the otic process
of the suspensorium.

The explanation of what is seen in these small kinds comes from the Menopome. In
that type, and in the Siren, there is an oval hyomandibular, over which the facial
nerve passes ; hut in the Menopome the columella also is very large, perfect, and
elegant.

In the Menopome, the gigantic Siebolclia, my figures (to appear in the Zool. Trans, for
1881) show this cartilage sticking by its broad foot to the top of the suspensorium and
squamosal above and behind, and having its narrow end inserted into a bony stapedial
sheath. This cartilage is formed beneath that fold of skin which is the operculum of the
first cleft. That cleft, in the Selachian, is only left open above ; in the Urodeles it is
closed ; but in the huge types just mentioned a " pharyngo-hyal " and an " epihyal "
appear ; the former becomes the " columella ;" in the Frog both elements are included
in that rod *.

The main difference between the fore part of the skull of the youngest larva of
Spclerpes salmonea and that of S. rubra (figs. 1-4) is the slender face and more advanced
condition of the specimen.

We have now the right and left moieties of the sphencthmoid (fig. 3) ; for close
behind the ethmo-palatine buds (e.pa) the skull-wall is ossified : these growths are still
soft. The internasal lamina? (i.n.c) are not so much advanced as in the last instance, but
the eth mo-palatines are larger.

The suspensorium, besides being more out-turned, has sent forth a small pterygoid
tongue of cartilage {e.pg) ; but this shows at present so little affinity for the bony plate
beneath it that it crosses it at a right angle. The cartilaginous process has at present
the same outward direction as the body of the suspensorium.

The investing bones and the inferior arches had nothing in them differing from the last
to make them noteworthy ; tbc more pointed skull of necessity has narrower roof- and
floor-bones.

Skull of larva of Spelerpes salmonea, 3| inches long.

These larva? were much longer than the last, they were just ready for undergoing the
Salamandrian metamorphosis, and yielded some very interesting results. One of the
most important of these is the slowness of these larva? in acquiring those characters
which would lift them up out of the level of Proteus and Ileuobranchus.

Large as these specimens were, I could detect no additional parts to the fore face ; they
were as simple there as in the types just mentioned.

There is not much at present to remark upon in the occiput ; the cartilage above is
less, but it is undiminished below ; and this specimen had more notochord (fig. 6, nc)
than the last.

* The account given of the columella of these Urodeles iu my "Abstract" (Linnean Society's Journal, Zoology,
vol. xiv. p. 719) is incorrect; it is not a " pseudo-columella ;" the facial nerve passes over it.

OF THE SKULL IN AMPHIBIA URODELA. 197

The odontoid vertebra (od.v) clearly shows its three elements, and these are all grafted
upon the " atlantal " or second joint (•» 1). The skull is here seen to be most carefully
articulated to the spine. The auditory capsules are a little more protruded outwards,
carrying with them the suspensoria.

The capsules are very large, owing to the great development of the three canals (fig. 5,
a.sc, h.sc, p.sc). The hinder ampulla gives a very swollen appearance to the capsules
even below (fig. 6). From this convexity we pass forwards to another, the vestibular ;
its ossification is advancing rapidly, leaving but little cartilage along the fenestra
(figs. 6 & 8). But the stapes and the window it fills are tilted now, so that they are only
seen obliquely in a basal view ; thus the handle, or stapedial rod (fig. 6, sp.c, fig. 8, co), is
getting nearer the top of the "tegmen." The relation of that rod to the facial nerve is
best made out by paring these parts away, and seeing them from within (fig. 8. 7,
pro, st). The nerve is imbedded in the prootic for some distance ; in fig. 8 its course
outside the bone is indicated by dotted lines up to the free edge.

The occipito-otic bone is spread into the base of the skull, ossifying the remains of the
parachordals behind and the stunted ends of the trabecular in front ; a notch shows the
division of these two regions. In front of the foremost basal lobe the bone ensheaths
the alisphenoidal wall for a short extent ; thence to the optic nerve (2) the wall is soft ;
but in front, nearly up to the small distinct semioval ethmo-palatine (e.pa), the sphe-
nethmoid (figs. 5-7, sp.e) has spread.

Directly in front of those bony tracts the trabecule, in their ethmoidal region, have
united ; there is a broad internasal lamina (i.n.c), gently emarginate behind and more so
in front.

The suspensorium is altered in some respects ; the pedicle (pd) is a more distinct
swelling ; the quadrate ossification (<?) is larger ; and the pterygoid cartilage {e.pg)* is not
only larger, but has also changed its direction ; it now runs parallel with the inner edge
of the pterygopalatine bone. That bone (figs. 5, 6, p.pff) has become inordinately large,
and the toothed region is now confined to the outer margin of the broad trilobate
anterior part. This bone lessens its breadth two thirds in the hinder half; it turns
outwards in a pedate manner, and is now applied to the lower face of the suspensorium,
almost reaching to the quadrate hinge.

The vomers also (v) are altered ; from their primary subcrescentic form (fig. 2) they
have now become, like those of the low Perennibranchs, straight, with an inner hook, the
hook and the outer margin only being dentigerous. Nearly the width of the ethmo-
palatine cartilages intervenes between the vomers and the pterygo-palatines.

The parasphenoid (fig. 6, pa.s), by its splintery fore edge, nearly reaches the middle
nasal hole (m.n.p) ; it only very gently increases in breadth up to the basitemporal
region ; it then is twice somewhat produced outwards, most at the first pair of pro-
jections, and ends behind, a little in front of the foramen magnum, by a slightly emar-
ginate edge.

* In fig. 6, PI. XX., this is lettered e. pa by mistake.
SECOND SERIES. — ZOOLOGY, VOL. II. 27

198 PROF. W. K. PAEKEE ON THE MOEPHOLOGY

Although nearly ripe for metamorphosis, there is but little promise in this stage of the
strange things that will appear in a few weeks afterwards.

The Skull of an adult Spelerpes rubra, 4| inches long.

This skull, all things considered, is one of the most remarkable I have seen; in
passing from the larval state to the adult, we seem to have climbed halfway up the
vertebrate scale. In its completeness, now that all the new things have been added,
it has become a singular commixture of the lower ichthyic, and the higher reptilian,
styles of cranial architecture.

In making this building grow, Nature has brought out of her treasures, and fitly
framed together, things new and old ; thus, well observed, this skull may teach us
much of what is possible in the morphological stages of a vertebrate cranium.

Having gone so far, if the Life-power of the individual had sufficed, such a skull, one
could imagine, might have gone on changing : a few steps in one direction would have
made it like that of the Emu, at the base of the Bird scries ; or in another, that of a
Monotrcmc, at the base of the Mammalian group.

The large pedunculated occipital condyles (PI. XXI. figs. 2, 3, oc.c) are wide apart,
and the posterior basicranial euiargination is perfectly filled by the curious intercalary
odontoid vertebra (ocl.v), which, being ankylosed to the next («'), gives the larger joint
an almost exact resemblance to the second (" odontoid " or " axis ") vertebra of an Ox.
Yet we have just seen that this bigger vertebra is the atlas to this skull, and that the
" odontoid process " is in reality a small postcranial vertebra, without an arch, and
segmented from the hind margin of the parachordal tract.

There is a considerable tract of cartilage, both above and below, in the occiput (figs. 1, 2) ;
for the rest, the occipito-otic mass, right and left, is continuous solid bone ; yet this
bony tract stops at the foramen ovale (5) ; for the ossification of the endocranium is not
intense in this type.

The auditory masses are very large and rugged ; the huge posterior canals di.sc), as
they turn outwards and backwards, have a sulcus separating them from the sharp edge
of the exoccipital, where it forms on each side a rim to the foramen magnum.

A very large triangular space exists between the emarginate occipital roof and the
front, almost transverse, margin of the " atlantal " vertebra.

Most of the anterior canal (a. so) is tiled over by the parietals (fig. 1, p) ; but the bony
matter projects from beneath the notched edge of those plates, forming a crest for the
origin of the fibres of the temporal muscle.

As the anterior canal is roughly covered by the parietal bone, so is the horizontal
(h.sc) by the supratemporal part of the squamosal (sq), which ploughs itself into the
substance of the prootic, raising it beyond its upper edge into a rough semicircular
ridge.

Mesiad of this ridge, the prootic is scooped for the temporal muscle ; thus there is a
crescentic fossa, with its convex margin outside ; it reaches to the large oblique balk
formed by the posterior canal.

OF THE SKULL IN THE AMPHIBIA UEODELA. 199

The under face of this ear-capsule is very different from that of most high " Uro-
delous " skulls.

In those massive typical Caducibranch skulls just described, e. g. Notophtlialmus
(PI. XVII.) and Cynops (PL XVIII. ), the vestibule is in the form of a smooth bulla,
with the stapes set on behind ; and that plate is either quite soft or very slightly ossified ;
it is also relatively small.

In Spelerpes (PI. XXI. figs. 2, 3, and PL XVIII. fig. 10) the vestibule is but little pro-
tuberant ; its fenestra is lateral, and, corresponding with the stapes, very large.

At the inner margin of the fenestra the bone is scooped, and partly unossified at the
edge. This is seen in the oblique view (PL XVIII. fig. 10). The outer margin of the
fenestra ovaiis is right under the tegmen, made larger and more beetling by the postero-
inferior edge of the squamosal (sq). This change of position is due to a sort of vegeta-
tive affinity or attraction between the stapes and the stapedial rod (st, co), which is found
in the inner layer of the temporal plane.

The 7th nerve (7) first burrows in and then grooves the prootic a good distance in
front of the stapes ; in its escape it curves a little backwards, and then turns obliquely
forwards and outwards under the otic process and its ossifying quadrate centre.

It will at once be seen that this is a very different relation of nerve and rod to what
is seen in the Batrachia, where the hyomandibular, mostly late in appearance, passes
under the facial nerve to apply itself, not to the outer face of the stapes, but along its
antero-superior margin, pushing into the scooped fossa that leads to the fenestra ovaiis.
Moreover, in the Batrachia the stapes never yields any bony matter to the columella ;
in them the stapes is seldom ossified ; and the columella generally cuts off the segment
that grows into the front of the fenestral fossa.

One more point of interest is to be noted here : the lowest Perennibranchs have their
stapes thoroughly ossified ; the highest Caducibranchs, like the Batrachia, have it soft ;
here, in Spelerpes, it ossifies early, and becomes a very perfect and elegant shutter
to that small oval window.

In the walls of the skull, from the foramen ovale (PL XXI. figs. 3, 5) to a small opening
behind the optic passage (2), the alisphenoidal region is soft ; then the wall is hard, up
to a small distance from the nasal sacs.

The nasal sacs (fig. 1, na) can be seen outside their roof-bones and also in front, forming
a rim to the external nostrils (e.n) ; but on the whole they are most carefully encased
in bone.

Yet in the last stage these olfactory paraneurals were not chondrified, nor even the
nasals, ectethmoidals, or maxillaries developed.

The condyle of the suspensorium reaches further forwards than the foramen ovale (5);
it is scooped in an angular manner (fig. 2), and above it the quadrate (q) grows upwards
to the pterygoid cartilage in front, and to the top of the otic process behind. But the
pterygoid cartilage and its base, running upwards as the ascending process to the alisphe-
noidal wall, is all soft.

In this thing Spelerpes is contrary to the ordinary Caducibranchs, namely, that

27*

200 PEOF. W. K. PAEKEE ON THE MOEPHOLOGT

instead of the pterygoid process of cartilage growing rapidly at first, and then, during
metamorphosis, becoming less and less, it comes late, grows slowly, and then does not
diminish in size.

This is a correlate of the curious abortion of the bony pterygoid which usually seizes
upon the cartilage, converting it into its own substance. This outgrowth of cartilage
is falcate, with the convex margin inwards, and the point looking towards the point of
the short zygomatic process of the maxillary.

The large pterygo-palatine bone is now reduced to two remnants, one in front and one
behind ; three fourths of it has been absorbed. The pterygoid piece (behind c.pg) is an
irregularly five-sided plate, applied to the inner face of the suspensorium in front and
below; it does not project forwards, but its edge there is concave; above it is only half
the width of the quadrate.

The foremost patches of teeth have crept off the vomers (v) on to the palatines, just
where these bones have coalesced, so that the exposed part of the vomer is edentulous.
Each palatine bone (pa) is mainly directed transversely ; it extends outwards much
further than the hinder process of the vomer, to enclose the inner nostril (i.n), and some-
what underlies the more extended ethmo-palatine cartilage (e.pa). Towards the mid
line it is curved backwards, lies beneath the parasphenoid (pa.s), has its concave facet
scooped, and soon ends in jagged projections.

Below, close behind the vomer, where the bone is the thickest, there is a row of sharp
decurved teeth, arranged in sigmoid manner. The greatest arch in this series is between
the internal nostril and the mid line of the skull; turning backwards, the right and left
rows nearly meet. Although the palatine bones go no further backwards, the tooth-
tracts do ; they are continued so far back as to end opposite the foramen for the facial
nerve (7). At first in a single row, and almost close together, they gradually widen their
plot until there are five or six teeth in one transverse line ; the interspace along the
middle of the parasphenoid also gently widens.

These lanceolate tracts, with their out-turned sigmoid fore end, are very elegant ; they
are the so-called sphenoidal teeth. These rasp-like tracts existed before the palatine
bones were formed, and grow far backwards, independently of them, where the bone has
been absorbed.

The vomers (fig. 2, v) are some distance apart, except behind, where they unite with
the palatines. Together they form a deeply notched fan, and each leafy plate has a
deep rounded notch on each hinder margin for the internal nostril {i.n). Their interior
margin is ragged, and their very large palatal development causes the palatine plates of
the premaxillaries and maxillaries (px, mx) to be narrow. Reference to the figures of
the last stage (PI. XX. figs. 5, 6, v) shows how strangely these bones have altered from
their larval form.

The premaxillary (px) is scarcely less transformed : looking at the bone in the three
larvae, we see it formed of a gently arcuate dentigerous part in front, and of two long,
lanceolate nasal processes that run back and lie over the frontals ; the middle nasal
passage (m.n.p) is seen between the roots of these crura and the internasal carti-
lage (i.n.c).

OF THE SKULL IN THE AMPHIBIA UEODELA. 201

Below (PI. XXI. fig. 2, px) there is no appreciable palatine belt behind the tooth-
bearing edge right and left, and the outer parts of the bone are fiat (figs. 1, 3); but in the
middle this bone is greatly developed vertically. Below (fig. 2) the two crura appear in
front of the vomers ; but between those bones they have coalesced ; thus an oval space is
enclosed (m.n.p). The coalesced part runs some distance backwards as a "median
palatine process ; " behind it is another space, over which lies the internasal plate of
cartilage.

Above (fig. 1, n.px) the original distinctness of the nasal processes is seen in front by a
deep hole, and behind by a sharp notch ; the two bars have grown into a very large
tract, which reaches back as far as the hind margin of the nasal sacs.

The middle nasal passage (m.n.p) is made into a deep well, with an everted funnel-
like mouth, which is so greatly lipped behind as to overshadow the solid continuation of
the bone, and give it the appearance of being a totally distinct plate. The nasals, also,
clinging to the well's mouth, hide the median part ; and it, in turn, becoming narrow
and splintery, is imbricated on the frontals (/), wedging them apart. The large nasals
(fig. 1, n) are roughly pentagonal ; they have a concave edge forwards, towards the
nostrils (e.n), and a sigmoid edge outside ; they project towards one another behind the
notched lip of the nasal well, getting under it. Behind, they run transversely, with irre-
gular teeth, imbricating the frontals on the outer part, but not reaching nearly so far
back as the premaxillary.

The ectethmoids (e.eth) are large conchoidal bones, that reach as far as the outer angle
of each nasal in front, and clamp the orbital edge of each frontal for some distance.

Each maxillary (fig. 3, mx) is like a hatchet, the blade being its facial plate, and the
handle its zygomatic process ; but the blade is cut away in front ; in that notch there is
a graniform septo-maxillary (s.mx).

To cover much of this long skull the frontals (/) are long ; they are widest in front,
are deflected over the cranial wall (fig. 3), but have little orbital development. Where
they begin to overlap the parietals (p) the skull is narrowest, and suggests the appearance
of the outline of an hour-glass.

The end of the frontals is more than halfway from the foramen opticum (2) to the
foramen ovale (5) ; the frontal suture is most irregular ; the edges of the bones being
very thin, and, as it were, lacerated.

The parietals (p) are much more elegantly finished. Each bone forms a wall-plate in
the interorbital region almost to the ectethmoid, and this part really becomes & paries,
as it descends with considerable steepness. The two bones gradually widen up to the
auditory masses, and then they suddenly attain their greatest width, not lessening much
behind, but ending transversely. Where they lie over the hind skull the parietals seem
to have been fitted on in a plastic condition, so thoroughly do they fit to each rising and
falling of the substance beneath. The dovetailing of the sagittal suture is perfectly
mammalian in its complexity ; the squamous sutures also, laterally, are very irregular
over the anterior canal (a.sc) ; the posterior canal (p.sc), where it joins the anterior, lifts
the lambdoidal edge of each parietal bone. The squamosals (sq) seem to grip the ear-
masses ; below, indeed, they are merely a thick wedge, binding the quadrate ; but above,

202 PEOF. W. K. PAEKER ON THE MOEPHOLOGY

the stipratemporal tract is oblong, curved, ridgy, and scabrous ; curling over the tegmen
tympani, it imbeds itself into the very substance of the bony roof of the horizontal
canal (h.sc).

From above (fig. 1, sq) this upper part looks like a distinct plate from the lower. The
side view (fig. 3) corrects this false impresssion.

The parasphenoid (fig. 2, pa.s) is not less elegant than the other parts ; its narrower
fore part is concealed by the palato-vomerine ankyloses ; thence it slowly widens, is much
hidden by the long tooth-beds, and then becomes winged behind.

The larger basitemporal projections are hooked ; then come lesser snags ; and behind,
near the foramen magnum, the bone is transverse and gently emarginate.

This bone is gently convex along its dentigerous part, but more so beneath the
medulla oblongata, where the endo-cranium is still deficient.

The mandible (figs. 3 & 4) is quite normal ; on the outside (fig. 3) is seen the den-
tary (d), nearly reaching to the angle, and the coronoid region of the articulare (ar) is
seen above it. The condyle is cylindroidal ; in front of it the still thick cartilage lies in
the trough of the articulare, which is not so long as in many kinds of Urodeles. The
splenial (sp) is a long, narrow, pointed splint, with but four teeth in old specimens ; the
teeth gradually become uniserial and are carried by the dentary.

The hyobranchial series of arches (fig. 5) have undergone their share of metamorphosis
(compare PI. XIX. fig. 9 with PL XXL fig. 5) ; and although several parts have been
absorbed, they are still a large and elegant structure. The hyoid arch is now perfectly
divided into a long, flat, curved, ceratohyal (c.hy) ; and a small terete hypohyal {h.hy),
which is loosely connected with its own main bar, and is turned backward to arti-
culate with the fore end of the first basibranchial (b.brx). The first branchial arch is
now composed of two pieces of equal length ; the lower piece (c.br1) is dilated and emar-
ginate below, and is attached loosely to its own basal bar (b.br1). The upper piece
(e.br]) is only dilated below, where it has received into itself the substance of the second
epibranchial ; both the ceratobranchials (c.br1, c.br2) articulate with it ; thus only the
lower segment of the second branchial arch is present now. This second ccratobranchial
is a slender rod, gently sigmoid, and articulates with the end of the first basibranchial
and the end of the first epibranchial.

The 1st basibranchial is broad and spatulatc in front, with a straight fore edge ; it then
narrows, and is elegantly enlarged behind for the facets of the 2nd ceratobranchial. The
main bar of the 2nd basibranchial (b.br1) has been quite absorbed, but its dilated hind
part exists as a large bony " thyrohyal ; " it is in the form of an unbent bow, the two
horns of which are united by a thick oval mass, the remnant of the longitudinal bar.

I am able, thanks to Prof. Mivart, to give an illustration of another adult of this
group of Caducibranchs.

Species 8. — The Skull of the adult Desmognathus fuscus.

This is another N. -American Newt of the same group as the last ; its skull differs only
in unimportant particulars, but some of these arc very instructive. This skull is alto-

OF THE SKULL IN THE AMPHIBIA UEODELA. 203

gether shorter and has a better finish, with denser and more polished bones ; it evidently
belonged to a more aged individual, as it has lost its sphenoidal teeth.

Tbe occipital condyles (PI. XXI. figs. 6-8, oc.c) are very large and pedunculated ; tbe
facets look more downwards than upwards ; tbe space between them is a measure of tbe
size of the intercalary (odontoid) vertebra. Tbe occipital arch is, on each side, entirely
confluent with tbe auditory mass ; above (fig. G) the arch is finished by suture, below (fig. 7)
by a narrow tract of cartilage.

Tbe auditory masses are much less rugged than in Spelerpes, tbe divisional fossa
and tbe swelling arches of the semicircular canals being more bidden by tbe more perfect
bony encasement ; and tbe parietals (p) are imbricated on the hind skull more smoothly
and with more exquisite finish.

Tbe cartilaginous tract between the prootics and the sphenethmoids (sp.e) is very small,
just a little rim to the bole for tbe trigeminal nerve ; but on the outer region above
there is a tract of cartilage where the otic process has coalesced with the capsule
(PL XXI. figs. 6, 8 and PI. XVIII. fig. 9, ot.p).

But below (PL XXI. fig. 7 and PL XVIII. fig. 9) tbe quadrate has run its bony sub-
stance into the apex of tbe process ; and between this part tbe horizontal canal (h.sc) is
completely encased in bone. The sulcus between the conjoined and the single canals
above (fig. 8, a.sc, p.sc, h.sc) is shallow and crescentic ; the squamosal (sq) docs not climb
on to this space nearly so much, but is more lateral.

Below (PL XXI. fig. 7 and PL XVIII. fig. 9) the fenestra ovalis is sublateral, but it
is less, as also its enclosed stapes (fs.o, st) ; that plate is roughly pyriform in outline, is
unossified at its rim, and has on its broad hinder part a boss for the muscular fibres that
are inserted in it. Also on the outer margin in front of the boss there is on the right
side a bony columellar handle, which is attached by ligamentous fibres to the under
surface of the " tegmen " (PL XXI. fig. 7 and PL XVIII. fig. 11, co, st)*. But on the
left side (PL XXI. fig. 7 and PL XVIII. fig. 9) this rod is smaller, and its hinder
end is attached by a ligament nearly its own length to the stapes ; the fibrous tissue is
continued from its fore end to tbe tegmen and otic process. This is the upper end of
the " suspensorio-stapedial ligament," such as is found in other Urodeles.

This curious onesidedness and irregularity, and the free development of this element
in the llenopome, leave no doubt in my mind as to its nature ; the facial nerve (7)
passes out beneath the otic process, further forwards, in front of this small columella.

The sphenethmoids (sp.e) reach from the rim of the foramen ovale to a short distance
from the nasal region ; they are fenestrate for the optic nerve (2), and they form a hard,
strong wall to this well-built little skull.

Only the narial rim (e.n) and the etbmo-palatine (e.pa) can be seen without dislocating
the bony plates, which are most artistically fitted and compacted together.

The condyles of the suspensorium are behind the foramina ovalia as much as they are
in front of it in Spelerptes ; the quadrate (q) ossifies up to the pointed apex of the otic
process (PL XVIII. fig. 9, ot.p), but the front of the suspensorium above is soft.

From the front a similar, but smaller pterygoid process (e.pg) runs forwards and out-

* In PL XVIII. both figs. 9 and 11 arc reversed.

204 PEOF. W. K. PAEKEE ON THE MOEPHOLOGY

wards, more outwards than in the last ; and although somewhat smaller, it comes near
to the larger maxillary spur {ma:). This pterygoid hlade has a base of cartilage, which
runs inwards to the alisphenoid region, and outwards into the fore part of the otic pro-
cess ; but behind, the quadrate (PI. XXI. fig. 7 and PI. XVIII. 9, q, ot.p) is seen as a
very solid wedge of bone, jammed closely in between the prootic and squamosal {pro, sq),
and flanked on its inner edge with the small pterygoid bone (pg). The condyle, carrying
a very solid mandibular ramus (figs. 7, 8, 9), is large, and has a deep and elegant troch-
lear cavity.

The pterygoid bone (pg) is still smaller than in the last ; it is a small ear-shaped plate
lying in the axil of the suspensorium, binding it on its inner face, and the prootic on its
anteroinferior aspect, just in front of the escaping facial nerve (7).

The other part of the palato-ptcrygoid bone {pa) is twice as large as the pterygoid ;
each plate is a small irregular hatchet, with a pointed blade that looks backwards ; the
back of the two blades are apposed, and each handle runs outwards and a little for-
wards, nearly completing the rim of the inner nostril (in), and partly underlying the
corresponding endoskeletal part, or ethmo-palatine (e.pa). The thickest part is where
the blade and handle join ; and there the bone is overlain by the vomer (v). These
bones, now, as well as the vomers, have lost their teeth, most probably, in this case,
through age. The whole plate is bound to its fellow by suture behind, and by simple
harmony in front ; and this structure is very remarkable in appearance, both the large
vomerine and the lesser palatine wings being separated by the pyriform internal
nostrils (i.n). The bones at the middle of the vomer, in one part of the suture, are
deficient, and leave an oblong membranous space beneath the septum nasi.

The palatine plates of the vomers are less than in the last kind, and those of the
jaws are larger (fig. 7, v, px, mx) ; that of the premaxillary can be well seen in front of
the vomers.

That bone is of less extent at its dentary margin than in Spelerpes, but its palatine
plate is more perfect, and the opening of the middle nasal cavity (m.n.p) is longer, both
above (fig. 6) and below (fig. 7).

The upper surface of the premaxillary (fig. 6, px) is smooth and even ; for although
the cavity in it is large and well defined, yet it is not wrought into a rimmed funnel, as
in Spelerpes ; yet the bone is sulcate behind the upper opening.

In front, on each side, the branches of the orbito-nasal nerve escape in the sloping
fore part of the bone at its solid part ; then a bony lamina is continued backwards,
enclosing the canal, and then widening, to end in two triangular processes that bind on
the fore part of the frontals (fig. 6, n.px,f). At its narrow part the premaxillary is
overlapped by the nasals («). These are very irregular triangles, with a concave base
looking forwards, and partly enclosing the cartilaginous narial rim (e.n) ; their outer
side lies on the maxillary (mx), and their apex on the corresponding frontals.

Contrary to wont, the ectethmoid (e.eth) is not seen outside the orbit ; whereas in
Spelerpes it has an unusually large external plate. In Desmognathus this is partly due
to the lesser size of the bone, partly also to the larger and more perfect facial plate of

OF THE SKULL LN THE AMPHIBIA TJEODELA. 205

the maxillary (mx). These latter bones unite freely with both the nasals and frontals ;
in Spelerpes they reach neither of these. The zygomatic process is also much more
developed, almost equal to what is seen in the Batrachians.

On the right side (PL XXI. fig. 6, s.mx) there is a very small septomaxillary ; it rests
upon the junction of the maxillary with the premaxillary. Thus there are two instances
of asymmetry in this skull ; for on the left side the small stapedial rod is, contrary to
wont, distinct.

The frontals and parietals (f,p) are exquisitely finished and smooth; yet there is but
little orbital development of either (fig. 8). Their sutures, the frontal and sagittal, seem
to belong to some much larger animal. They are like what we see in Mammals ; but the
dovetailing projection and notches are out of all proportion to the size of the bones thus
interlocked.

The lambdoidal suture is squamous and almost transverse.

The squamosal (sq) is more ichthyic than in the last ; it appears but little on the upper
surface (fig. 6), and rather serves above as a buttress to the ear-mass set under the
" tegnien." It is a short wedge, thick above and roughly pointed below.

The parasphenoid (figs. 7, 8, pa.s) is more primordial in form than in Spelerpes ; it is
oblong in the long interorbital region, has a short triangular ascending basitemporal
process, and under the medulla oblongata it is almost straight-sided again, but wide ; it
ends subtransversely, with a bevelled crenate margin, a moderate distance in front of the
foramen magnum.

The mandible (figs. 8, 9) is very stout and solid, and shows why the suspensorial
condyles should have been so well scooped. The dentary (d) reaches nearly to the end of
the ramus : it has a very high and thick edge at the chin for its fellow ; it has coalesced
with the splenial (fig. 9, sp), and has overlapped and partly absorbed the articulare (ar)
and the Meckelian rod. The articulare has largely ossified the cartilage, leaving, how-
ever, a solid cylindroidai condyle ; it sends upwards a strong rounded coronoid process
in front of the condyle.

The hyobranchial series is much like that of Spelerpes ; it is still slenderer, and equally
unossified. The ceratohyals (fig 10, c.hy) are quite Like those of the last ; but the
hypohyals (h.liy) are stouter, and are directed forwards and inwards, instead of the
contrary direction seen in Spelerpes. The first cerato-branchial (c.br1) is lanceolate,
instead of being dilated ventrally. The first epibranchial (e.br1) is still slenderer than in
the last ; it is not dilated to receive the next rod, which is merely applied to its inner
side. The first cerato-branchial lies against the basal piece, as in the last, and the
second articulates with its end ; but this second bar (c.br2) is longer than in Spelerpes,
and instead of being bent forward is bent backwards.

The bony remnant of the second basibranchial (b.br2) is only half as large as in the
last. It has two projections at the middle. Unlike the last, it is altogether bony, having
no soft cartilage tipping its horns.

SECOND SERIES. — ZOOLOGY, VOL. II. 28

20G PEOF. W. K. PARKER ON THE MORPHOLOGY

Summary and Conclusion.

These instances of the Structure and Development of the Skull in the Caducibranchiate
division of the Urodeles are intended to be a continuation of my recently published
account of the skull in the Tailed Amphibians in the Philosophical Transactions. But
I am only in the midst of things at present ; and this paper must be followed by one
giving the development of the skull in Triton, the structure of the adult skull in
the Menopome and Siren and Amphiume and the skull of an old Sieboldia.

When time has permitted all these things to appear, and when my fast-growing materials
relative to the Anurous type of skull have seen the light (Phil. Trans, pt. 1, 1881), then it
will not be difficult to dilate upon the growth and architecture of the skull in these most
instructive metamorphosing " Ichthyopsida." Meantime a narrower scope will be of
some use ; for the tail-bearing forms undergo great and important changes during their
life-history, and in them we seem to see low and ancient types of " Dipnoi " making trial
of land-life, and becoming, in very various degrees, adapted to their new surroundings.

On the other hand, the Batrachia appear to be the terrestrial outcome of low and
generalized tribes, which were not Marsipobranchs, nor Chimaeroids, nor Selachians proper,
but all these in potency if not in specialization. The few bony plates that in the
Dipnoi become attracted to the chondrocranium, and wedded to it — these appear early in
the life of each individual Urodele.

These, and scarcely any others, are the heritage of the lowest Perennibranchiate forms,
as Proteus and llenobranchus ; and every larval Urodele is like the fishes at first, and
only after a time gains the additional parts that characterize the most developed types
of the group.

Some of these almost rival the more ambitious tailless Amphibia ; they form a sort of
parallel " leader " in the vertebrate life-tree, but more simple and more stunted.

Metamorphoses so wonderful as these Amphibia are seen to undergo (they transform
themselves openly, and not in secret places, as is the wont of most Vertebrates), render
them invaluable to the student ; for they train the mind, and make it much more easily
take the stamp of morphological teachings.

The Urodeles are inferior to the Batrachia in that they do not often metamorphose
their hyomandibular into a columella. This cartilage, or " pharyngo-hyal rod," never
gains an annulus tympanicus ; for their first cleft is not only closed, it was never
open, and was arrested early.

One type has a simple hyomandibular, viz. Proteus ; but it remains unchanged, as in
Ceratodus, Notidanus &c.

This type of skull differs from that of the Batrachia in that it never acquires a carti-
laginous floor, and its sides, ossifying, do not unite to form a girdle-bone.

The pterygopalatine arch, such as we see in the Batrachia, is always formed in the
Urodeles of two totally separate elements, which never unite ; as in the Skate, there is a
distinct " ethmo-palatine " and a pterygoid process of the suspensorium. Moreover, the
bony plates that are applied to these two elements begin irrelatively to them, as the
homologue of the archaic pterygo-palatine submucous plate of the Dipnoi, by subdivision

OF THE SKULL IN THE AMPHIBIA UEODELA. 207

and curious metamorphosis, both in form and in position, this bone yields the ossifyin°-
elements to each of these diverse and never united parts. Instead of repeating the
" labials " of the Lamprey and its relations, they have, both in their larval and adult
condition, the " median nasal cavity." Also, having no labials to metamorphose, they
have no superadded narial valves, and no ossified mento-Meckelian pieces in their man-
dibles.

The premaxillary is often azygous. They have no " quadrato-jugal " binding the
maxillary to the quadrate ; they may have a small jugal, and a small additional tem-
poral ; both these exist in Amblystoma opacum.

Equally with the Batrachia they have, as a rule, in the adult, no supra- nor basi-occi-
pital ; but two rudiments of occipital centra may exist for a few weeks of larval life, as
in Seironota perspicillata.

Again, the Urodeles differ from the Batrachia in developing the curious intercalary
vertebra, the " odontoid " pre-atlantal rudiment.

Their basal palatine, their palatine and sphenoidal teeth, their dentigerous splenial,
their distinct hypohyal, the formation of their thyrohyal from the dilated end of a median
branchial piece — all these things distinguish this lower tail-bearing group from the
higher tailless section of the Amphibia.

Finally, there are curious dovetailings of the characters of the two groups : for
instance, there is usually no ossified quadrate in the Batrachia; it is exceptionally
unossified in one of the " Urodeles " {Siren lacertina).

The divided state of the girdle-bone is seen in the anourous Pseudophryne Bibronii
and other Australian types ; these also have a rudimentary occipital centrum, or basi-
occipital.

Pipa, like the Urodeles, has no quadrato-jugal, and has the quadrate ossification.

The Bombinator (B. igneus) and Psexidophryne Bibronii have no columella; the
former and Pelobates fuscus have no " annulus : " all the " Anura" want this for a
time ; it appears at very variable periods in tbem. In the Urodeles the trabecule do
not, as a rule, develop a secondary cornu (prorhinal) from the inner angle, such as is
seen in Batrachia and embryo Sharks.

APPENDIX.

Since the above descriptions were written (latter part of 1877) my first paper on the
Urodelous type of skull has appeared in the Philosophical Transactions (1877, pis. xxi.-
xxix. pp. 529-597). Of my own finished work there still remains for publication after
this the development of the skull in the Common Newts (Triton cristatns, Lissotriton
punctatus), the skull of the adult Menopome, and of the half-grown and adult Siren
lacertina. But the MS. for the present communication was no sooner finished than
I received from Dr. Robert Wiedersheim, of Ereiburg, his two invaluable Memoirs on
these types.

28*

208 PEOF. W. K. PAEKEE ON THE MOEPHOLOGY

I. The first of these is on the anatomy of Salamandrina (Seironota &c.) perspicillata,
and Geotriton (Besmognathus) fuscus (Genoa, 1875). This exquisite work contains
figures and descriptions of several other kinds besides these two. In my paper just
referred to, I give the skull of both the larval and adult Seironota perspicillata.

II. The second work is entitled ' Das Kopfskelet der Urodelen, ein Beitrag zur
vergleichenden Anatomie des Wirbelthier-Schadels ' (Leipzig, 1877).

III. A third Memoir has now appeared (1878) by the same excellent anatomist, viz.
"Zur Anatomie des Amblystoma JFeismanni" (Zeitsch. f. wiss. Zool. Bd. xxxii.
taf. xi., xii.).

In the second of these important works there are not only figures and descriptions of
the skulls of adults of many of the kinds worked out by me, but also several that I have
not been able to obtain. These are as follows : — 1. Amphluma tridactylum, 2. Crypto-
branchus japonicus* , 3. Ellipsoglossa ncevia, 4. Manodon sibiricus, 5. Plethodon glutinosus,
6. Batrachoseps attenuatus, 7. Gyrinoph'dus porphyriticus, 8. Anaides lugubris, 9. Spejerptes
fuscus, 10. Triton subcristatus, 11. Triton platycephalus, 12. Salamandra atra, 13. Am-
blystoma punctatum. In this invaluable work figures are given also of the skulls of
Triton (Notophthalmus) viridescens, Besmognathus fuscus, and Triton torosus (Taricha
torosa), which are also figured and described in the present communication.

IV. It is also proper to mention here a Memoir by Dr. Philip Stohr (Prosektor
am vergleichenden anatomischen Institut zu Wiirzburg), entitled ' Zur Entwickelungs-
geschichte des Urodelenschadels ' (Leipzig, 1869).

There are very few things in these works that I see occasion to criticise, but very
much indeed to praise ; but there is one thing that must be noticed as an error in the
second work, namely, the cartilage attached to the stapes in the skull of the Menopome
(pi. ii. figs. 24, 25, Op) is much too small ; it was evidently drawn after it had become
shrunken. In the same plate (fig. 16) an error of my own is corrected (" Skull of
Urodeles," pi. xxvii. fig. 6) which failed to give a " splenial " to the mandible of Proteus
anguinus. At the least these works double our knowledge of the adult skulls in these
types ; and with these works before him, my published paper in the ' Philosophical
Transactions,' and the plates and descriptions now given, the reader will be able to form
a very good idea of the Urodelous type of skull, its modifications, and its development.
I shall, however, to help to this end, make a few comments on what is displayed in the
plates of Prof. Wiedersheim's second work.

Plate i. fig. 1. This figure of the skull of Menobranchus shows a fenestrated car-
tilaginous nasal roof (N.K, Nasenkapsel). Prof. Huxley (Proc. Zool. Soc. March 17th,
1870, plate xxx. fig. 1, p. 197) does not give this.

Plate i. figs. 11, 12, M.F. The hyomandibular is shown in Siren; the author
calls it " hinterer, mit dem Hyoidbogen durch ein Ligament verbundcner Fortsatz des
Quadratum." This epihyal element is partly confluent with the back of the suspen-
sorium.

Plate ii. figs. 21, 22, and figs. 24, 25, H.F, shows the same part in Cryptobranchits

* Worked out siuco.

OF THE SKULL IN THE AMPHIBIA URODELA 209

and Menopoma. I have worked it out in Siren, Sieboldia, and Menopoma, and am con-
fident of its real nature ; it is twice as large, and not confluent, in Proteus.

Plate ii. In fig. 22, Op, a stapedial stem is figured in Cryptobranchus as well as in
Menopoma (figs. 24, 25). This is the curiously modified pharyngo-hyal cartilage and
stapes, confluent.

Plate v. In Ellipsoylossa (fig. 65), in Banodon (figs. 69, 70), and in Amblystoma punc-
tatum (figs. 76, 77), the hyomandibular {II. F.) is shown ; in Banodon there is evidently
an additional stem to the stapes {Op).

The antorbital (" ethmo-palatine ") cartilage is shown (A.F.) to he confluent with
the pterygoid process of the suspensorium {Ftc) in Banodon (pi. v. figs. 69, 70), a
most important modification; for in the Batrachia that is a primary state of things,
which here is manifestly caused by confluence of the parts.

In the Toads (Bufo vulgaris, pantherinus, &c.) the subdivision of this part is secondary,
both of the palatine from the pterygoid and of the former from the trabecula.

In pi. vii. fig. 94 another most important exceptional condition is shown in Batra-
choseps, which is common enough in the Batrachia, viz., a large fontanelle in the roof of
the skull, arising from the deficient growth of the frontals and parietals.

In pi. i. fig. 8 (Bs.H, U.p.H?) the hyoid of Amphiuma is shown to be broken up,
distally, into even more pieces than I have found in Menopoma, where each hypohyal
is composed of three cartilages. A curious thing also occurs in Osseous Pishes, where
the hypohyal segment, quite distinct, is ossified by two centres.

The figures of transversely vertical sections of adult skulls and the various modifica-
tions of the hyobranchial arches in the adult are beautifully perfect, copious, and
instructive.

The remarkable {Petromyzine) " cavum intermaxillare resp. nasale (CI) is shown in
most of the types. I know of nothing in the morphology of the skull more tantalizing
than these structures. Surely some low ichthyic types will help us to connect this
structure with the azygous nose of the Lamprey *.

* Dr. Wiedersheim's description of the skeleton of Pleurodeles Waltlii (Alorphologischo Studien, plate ii. figs. S— 10,
pp. 27-42, Jena, 1880) may now be added to the above.

210

PROF. W. K. PARKER ON THE MORPHOLOGY

DESCRIPTION OF THE PLATES.

The abbreviations used in the different figures are as follows

al.s. Alisphenoidal region.

a. p. Ascending process of sus-

pensorium.
ar. Articulare.

a.sc. Anterior semicircular canal.
a.tr. Apex of trabecula.
cm. Auditory capsule.
au.c. Auditory capsule (or cavity) .

b.br. Basibranchial.
b.hy. Basiliyal.

b.o. Basioccipital region (or rudi-
ment) .

c.br. Cerato-branchial.
c.hy. Cerato-hyal.
co. Columella.
c.tr. Cornu trabecular.

d. Dentary bone.

e.eth. Ectethmoidal bone.

e.n. External nostril.

e.o. Exoccipital.

ep. Epiotic.

e.pa. Ethmo-palatine cartilage.

e.pff. Epiptcrygoid cartilage.

eth. Etbmoid.

/. Frontal bone.
f.m. Foramen magnum.
fo. Fontanelle.
fs.o. Fenestra ovalis.

h.hy. Hypobyal.
h.sc. Horizontal
canal.

i.n. Internal nostril.
i.n.c. Internasal cartilage.

w.

Investing mass.

semicircular

Ix. Larynx.

rn.h.l. "Mandibulo-byoid ligament.
mk. Meckel's cartilage.
m.n.p. Middle nasal passage.
mx. Maxillary bone.

n. Nasal bone.

na. Nasal roof-cartilage (nasal

" paraneural ").
nc. Notocbord.
n.px. Nasal process of premax-

illary bone.

oc.c. Occipital condyle.
od.v. Odontoid vertebra.
ol. Olfactory capsule.
o.s. Orbito-spbenoidal region.
ot.p. Otic process of suspen-
sorium.

/;. Parietal bone.
pa. Palatine bone.
pa.s. Paraspbenoid bone.
pd. Pedicle of suspensorium.
pff. Pterygoid bone.
p.pa. Posterior palatine bone.
p.pff. Palato-pterygoid bone.
p.px. Palatine region of pre-

maxillary bone.
pro. Prootic bone (or region) .
p.sc. Posterior semicircular canal.

px. Premaxillary bone.
py. Pituitary body-space

or

region.

q. Quadratum (bone or region of) .

s.mx. Septo-maxillary bone.

s.o. Supraoccipital region.

sp. Splenial bone.

sp.e. Sphencthmoidal bone.

sp.t. Spbenoidal teeth.

sq. S(juaniosal bone.

st. Stapes.

t.cr. Tegmen cranii.

tr. Trabecula.

t.ty. Tegmen tympani.

v. Vomer.

1.

Olfactory nerve.

2.

Optic nerve.

5.

Trigeminal nerve.

51

First or orbito-nasal branch

of trigeminal nerve.

5=

, 53. Upper and lower maxil-

lary branches of trigeminal

nerve.

7.

Facial nerve.

7'

. Palatine ("Vidian") branch

of facial nerve.

8.

Auditory nerve.

9.

Glosso-pharyngeal nerve.

i).

Vagus nerve.

Plate XIV.

Fig. 1. Skull of an embryo of Salamandra maculosa, three-fourths ripe (1st Stage), upper view,
x 17 diameters.

2. The same, lower view, x 17 diam.

3. The visceral arches of the same skull, upper view, x 17 diam.

4. Vertically transverse section through fore part of nasal region of a similar skull, x 17 diam.

OF THE SKULL IN THE AMPHIBIA UKODELA. 211

Fig. 5. Another section taken further back, x 17 diam.

6. A section of the same skull through the middle of the auditory capsule, x 17 diam.

7. Hyohranchial arches of adult (3rd Stage) , one side partly removed, x 5 diam.

Plate XV.

Fig. 1. Salamandra maculosa. Another section (1st Stage), through the eyeballs, x 17 diam.

2. A similar section made behind the eyeballs, x 17 diam.

3. A section made through hinder part of auditory capsule, x 17 diam.

4. A section made at junction of exoccipital region with ear-capsules, x 17 diam.

5. Skull of a ripe embryo of same species (2nd Stage), upper view, x 15 diam.

6. The same, lower view, x 15 diam.

Plate XVI.

Fig. 1. Skull of the adult Salamandra maculosa (3rd Stage), upper view, x 5 diam.

2. The same, lower view, x 5 diam.

3. The same, side view, x 5 diam.

4. Chondrosteous endocranium of same, upper view, x 5 diam.

5. Vertical section, with investing bones, showing interior of skull, x 5 diam.

Plate XVII.

Fig. 1. Skull of Notophthalmus viridescens (larva), 1 in. 28 lines long, upper view, x 15 diam.

2. The same, lower view, x 15 diam.

3. Part of skull of a similar larva, upper view, x 15 diam.

4. Visceral arches of same, upper view, x 15 diam.

5. Skull of adult of same species, upper view. x 7^ diam.

6. The same, lower view, x 7\ diam.

7. The same, side view, x 7\ diam.

8. Mandible of same, inner view, x 7\ diam.

9. Hyohranchial arches of same, with larynx, upper view, x 7% diam.

Plate XVIII.

Fig. 1. Skull of adult Cynops 2»jrogaster, upper view, x 6§ diam.
x 2. The same, lower view. X G§ diam.

3. The same, side view, x 6| diam.

4. Mandible of same, outer view, x Gf diam.

5. The same, inner view, x Gf diam.

6. Front view of skull of same, x 13] diam.

7. Hyohranchial arches of same, x Gf diam.

8. Part of auditory capsule and stapes of second larva of Spelerpes salmonea*. x 22i diam.

9. The same part in the adult skull of Desmognathus fuscus, left side, x 20 diam.

10. The same part of the skull in the adult Spelerpes rubra, x 20 diam.

11. Stapes of Desmognathus fuscus, right side, x 25 diam.

Plate XIX.

Fig. 1. Skull (part) of Omjchodactytus, adult, upper view, x 10 diam.
2. The same skull (perfect), lower view, x 10 diam.

* Inadvertently this is given as X 30 in the plate itself; and the figs. 9 and 10 have been wrongly numbered in
the plate.

212 ON THE MORPHOLOGY OF THE SKULL IN AMPHIBIA FRODELA.

Fig. 3. Visceral arches of same, part of upper view, x 10 diam.

4. Skull of adult Taricha torosa, upper view, x 10 diam

5. The same, lower view, x 10 diam.

6. Mandible of same, outer view, x 10 diam.

7. The same, inucr view, x 10 diam.

8. Hyobranchial arches of same species, x 4 diam.

9. Visceral arches of larva of Spelerpes rubra, upper view, x 13 diam.

Plate XX.

Fig. 1. Skull of larva of Spelerpes rubra, 1 in. 10 lines long, upper view, x 15 diam.

2. The same, lower view, x 15 diam.

3. Skull of 1st larva of Spelerpes salmonea, 2 in. 1 line long, upper view, x 10 diam.

4. The same, lower view, x 10 diam.

5. Skull of 2nd larva of Spelerpes salmonea, 3f in. long, upper view, x 7£ diam.
G. The same, lower view, x 7h diam.

7. Side view of part of same, x 7^ diam.

8. Part of auditory capsule of same, showing inner face of stapes, x 22^ diam.

9. Part of auditory capsule of Spelerpes rubra (larva), lower view. x 30 diam.
10. The same in 1st larva of Spelerpes salmonea. x 30 diam.

Plate XXI.

Fig. 1. Skull of adult Spelerpes rubra, upper view, x 5 diam.

2. The same, lower view, x 5 diam.

3. The same, side view, x 5 diam.

4. Mandible of same, inner view, x 5 diam.

5. Hyohranchial arches of the same, x 5 diam.

G. Skull of adult Desmnynathusfuscus, upper view, x 6 diam.

7. The same, lower view, x G diam.

8. The same, side view, x 6 diam.

9. Mandible of same, inner view, x G diam.
10. Hyobranchial of same, x 6 diam.

[ 213 ]

VI. On the Tusks of the Fossil Walrus, found hi the Bed Crag of Suffolk. By E. Rat
Lankester, M.A., F.B.S., F.L.S., Professor of Zoology and Comparative
Anatomy in University College, London.

(Plate XXII.)

Bead 6th May, 1880.

IN the year 1S65 I described, iu a memoir published by the Geological Society of
London, the tusks of a large mammal closely resembling, if not identical with, those of
the living Walrus. Several large fragments of these tusks, obtained from the base
of the Suffolk " Crag," had come to my knowledge, when, in the summer of 1861, 1 was
shown a small fragment of a similar tusk by Professor P. J. van Beneden at Louvain.
This fragment had been obtained from the Antwerp Crag ; and at the time of my visit
Professor Van Beneden did not indicate to me what conclusion he had formed as to its
nature. Subsequently, in the autumn of 1861, Professor Van Beneden communicated
in a letter to me his conclusion that the fragment of a fossil tusk in his possession had
belonged to a "Walrus-like animal, for which he informed me that he intended to use
the generic term Trichecodon.

Having myself in the meantime ascertained that the Suffolk specimens presented
all the essential structural features of the Walrus-tusk, I proceeded to figure and
describe them, adopting, with acknowledgment, the generic name suggested by Prof.
Van Beneden, and supplying a specific title : I named the fossil Walrus of the Suffolk
Crag Trichecodon Ruxleyi. At the same time I may point out that I was unable
to adduce any character in support of a generic distinction between the Walrus now
inhabiting the polar sea and that which lived in the Miocene period. It was therefore
only in response to Prof. Van Beneden's suggestion that the genus Trichecodon was
proposed ; and I am anxious now to withdraw my use of that term, and to refer
the Suffolk tusks to the genus Trichecus simply. The fossil Walrus of the Crag,
indicated by the fragments of tusks already made known, and by other more perfect
specimens to be described below, will therefore stand simply as Trichecus Huxleyi
(Lankester, sp., 1865).

I am glad to take the present opportunity, when I am about to submit some
excellent drawings of two nearly complete tusks from Suffolk, of clearing up, as far as
I am concerned, a certain confusion which has arisen in connexion with the nomenclature
of the fossil Walruses of the Pliocene deposits of Suffolk and Belgium. A splendidly
illustrated memoir on the fossil remains of marine Carnivora obtained from the environs
of Antwerp has recently been published by Professor Van Beneden (' Annales du Musee
Royal d'Histoire Naturclle de Belgique,' tome i. 1877). In this work a large number of
bones belonging to various Walrus-like individuals are figured, and are referred by
Prof. Van Beneden to two genera, Trichecodon and Alachtherium. I am not anxious on

SECOND SERIES. — ZOOLOGY, VOL. II. 29

214 PKOF. E. EAY LANKESTEE ON WALRUS-TUSKS

the present occasion to discuss the value of the characters by which Prof, Van Beneden
considers himself justified in founding two genera — the one, Alachtherium, having been
previously proposed without any definition by the Viconite du Bus (Bulletin de l'Acad.
Boy. Belg. 1867, p. 562), whilst the other had been adopted by me for the Walrus-like
tusks in 1865. But I must point out that though Prof. Van Beneden figures a number
of bones of his Walruses, he is unable to bring any evidence with regard to the tusks
beyond that afforded by the small fragment which I saw in 1861. There is actually
no evidence whatever that the particular bones referred by Prof. Van Beneden to the
genus Trichecodon have any thing to do with the fragment of a tusk in his possession ;
certainly there is nothing to show that these bones have any special relationship to the
Suffolk tusks to which iu 1865 I applied the generic term Trichecodon. As applied to
bones, ''Trichecodon," is an obvious misnomer; and there appears to be no reason
whatever for assigning that name to the bones to which it has been assigned by
Prof. Van Beneden, rather than to the bones to which he has applied Du Bus's generic
term "Alachtherium." The genus Alachtherium was indicated by Du Bus in 1867.
The first published reference to Trichecodon by Prof. Van Beneden is in 1871 (Bulletin
de l'Acad. Boy. Belg. 2nd ser. t. xxxii. p. 164), where he cites both bones and the
solitary fragment of a tusk (which is all as yet recorded from Belgium), and applies to
these remains the title Trichecodon Koninckii.

It must be obvious that, in default of specimens showing both bones and tusks in
juxtaposition, it is perfectly hopeless to attempt to identify either Prof. Van Beneden's
own fragment of a tusk or the Suffolk specimens with those bones which he calls Alach-
therium on the one hand, or with those which he calls Trichecodon on the other. At the
same time, should there really be only one Walrus-like animal proper to this period,
neither Alachtherium Cretsii of Du Bus (1S67), nor Trichecodon Koninckii of Van
Beueden (1S71), has priority as its title, but Trichecodon Huxleyi, mini (1865). Further,
if it should appear (as I am inclined to think it will) that there is insufficient ground
for the generic subdivisions indicated by the terms Alachtherium and Trichecodon,
the species Trichecus Huxleyi will yet have precedence, and any bones which may be
referred to the same species as that which carried the tusks which are so well represented
in Suffolk will have to bear the title T. Huxleyi.

This being the case, the following passage extracted from Prof. Van Beneden's
memoir appears to me to be open to correction on one or two important points. He
says, p. 53, under the heading Alachtherium Cretsii, " On a trouve en Angleterre
d'enormes dents dans le crag que Ton avait meme attributes au Dinotherium et qui pro-
viennent probablement de l'animal qui nous occupe. Une dent canine est figured dans

les Proceedings de laSociete Geologique de Londres de 1865, pi. xi En 1S65,

M. Ray Lankester fait mention de ces dents a la Societe Geologique de Londres ; mais
[A] comme on ne connaissait que le genre que j'avais £tabli sous le nom de Trichecodon,
il avait cru devoir les attribuer a une espece nouvelle, qu'il d6dia a M. Huxley, Triche-
codon Huxleyi Mais comme nous avons deux genres fossiles differents de

Trichecides, V Alachtherium et le Trichecodon, [B] auquel des deux ces grandes defenses
appartiennent-elles ? [CJ Nous pensons que e'est a Y Alachtherium, qui s'eloigne plus

FROM THE EED CRAG OF SUFFOLK. 215

du morse que Pautre, et avec d'autant plus de raison que [D] uous possesions des
defenses de Trichecodon qui out uu tout autre caractere."

In the first place, with regard to the passage marked A, it has escaped Prof. Van
Beneden's memory that he had not established the genus Trichecodon in 1865, but
suggested the name in a letter to me as one well fitted to indicate the fossil tusks of a
Walrus-like animal. Accordingly there was no thought of any genus or species
founded by Prof. Van Beneden when Trichecodon Suxleyi was described by me, for the
very sufficient reason that no such genus and no species of it had been described at that
time by the learned anatomist of Louvain.

With reference to the question marked B, there is no room for doubt, since the genus
Trichecodon was instituted by me for the very tusks in question.

In the sentence marked C, the Professor expresses an opinion which is entitled to all
the weight which his authority can give. He considers that the Suffolk tusks should be
connected with the bones to which he applies Du Bus's generic title Alachtherium, for
the reason that these bones are less like those of the living Walrus than are the second
series, to which he assigns the name Trichecodon. If his opinion were correct, the genus
Alacktherium (18G7) would have to give way to the earlier published name Trichecodon
(1865). But I cannot admit that there is the remotest evidence for the connexion of the
Suffolk tusks with the bones called Alachtherium rather than with any other Walrus-
bones, since no Walrus-tusk has been found in connexion with Walrus-bones in either
Suffolk or Belgium.

Finally, Prof. Van Beneden states (in D) that he possesses tusks of his Trichecodon
(1871) which have quite a different character from those of my Trichecodon (1865). He
does not figure these tusks nor elsewhere describe them. I find no reference to them in
his description of Trichecodon Koninclni. No tusk was obtained in relation with the
bones which he describes under that name. The single fragment of a tusk, which is the
only tusk at present recorded (except in the passage just quoted) from the Belgian Crag,
presents no character which can separate it from the Suffolk tusks, or lead to its iden-
tification with any particular Walrus-bones.

Further, supposing that there were such evidence forthcoming, it would be necessary
not to refer my Trichecodon Suxleyi to some new genus, but to create a new designation
for Prof. Van Beneden's later described form, and thus to avoid applying to it a desig-
nation already in use.

My conclusion, therefore, from a careful consideration of Prof. Van Beneden's state-
ments in his large monograph (1877), and from that of his shorter memoir (1871),
and of Du Bus's account of Alachtherium (1867), is that there is no evidence for the
association of the tusks of Trichecus (Trichecodon) Suxleyi of Suffolk with any one
set of the bones of Walrus discovered at Antwerp rather than with any other ; and
inasmuch as the tusks which we now possess furnish as sound a basis for generic and
specific characterization as do detached and fragmentary bones of the general skeleton,
the title Trichecus Suxleyi should hold its place ; whilst, further, if the generic term
" Trichecodon " is to be used at all, it is applicable not to bones which give no specific

29*

216 PEOF. E. EAT LANKESTEE ON WALEUS-TUSKS

information relative to the teeth, but to the teeth themselves, in the sense in which
I made use of it fifteen years ago at Prof. Van Beneden's suggestion.

New Specimens of the Tusks of Trichecus (Trichecodon) Huxleyi, and Comparison of
these with the Tusks of Trichecus rosmarus.

Two of the finest collections of the fossil bones and teeth derived from the base of the
Suffolk " Crag," viz. that of Mr. Whincopp of Woodbridge and that of Mr. Baker of
Woodbridge, have been broken up and dispersed during tbe last few years ; on the other
hand, the Ipswich Museum, which already possessed some valuable specimens, has been
enriched by the acquisition of the collection of the Rev. Henry Canham, purchased and
presented by Sir Bichard Wallace, Bart., and by specimens presented by Mr. Packard,
of Birkfield. The Ipswich Museum now contains the finest series of these remains in
any one collection. The two large tusks drawn in PI. XXII. figs. 1 & 2, formed part of
Mr. Canham's collection. They are the finest specimens which have been discovered,
being nearly complete tusks, and were kindly lent to me by Dr. Taylor, F.L.S., the
Curator of the Ipswich Museum, for study and illustration.

Prom an examination of the large series of tusks of Trichecus rosmarus preserved in
the Museum of the Boyal College of Surgeons (for the opportunity of studying and
measuring which I am indebted to Prof. W. H. Blower, F.B.S.), I have been able to
form some conclusions relating to the absolute size, sectional diameter, curvature, fluting,
attrition, and similar points exhibited by these great canines at different ages and in the
two sexes.

Such a series of Walrus-tusks as that in the Museum of the Royal College of Surgeons
exhibits a very great variety of forms and sizes ; and the first point to which I have to
draw attention is that the variation is precisely parallel to the variation met with in
the fossil Walrus-tusks of the Suffolk Crag. The chief differences of form which I have
been able to recognize are connected as follows with age and sex : —

1. Small tusks, almost straight, with unworn points and large pulp-cavity. These
belong to young individuals, and may differ a little according to sex.

2. Pull-sized tusks, more slender and more curved than No. 3. The fluting or ridges
on the flattened sides of the tusk much less pronounced than in No. 3. These appear to
belong to the females of Trichecus rosmarus.

3. Pull-sized tusks not longer than No. 2, but less ciuwed and more massive, having
a larger transverse diameter and a more decided grooving and ridging of the flattened
sides of the tusk. Some of the tusks of this type are nearly straight ; others are slightly
curved ; but they appear never to be so much curved as are No. 2.

•4. Short massive tusks, with the pulp-cavity filled by osteodentine. These are the
tusks of old individuals worn down, and exhibit a difference in their transverse section
and bulk, according as they have been derived from male or female.

The differences presented in this series of recent Walrus-tusks is paralleled by those
of T. Huxleyi from the base of the Suffolk Crag.

We have not unfrequently small specimens from the Suffolk Crag corresponding to
Group 1 of the recent tusks — that is to say, little or not at all worn, but very slightly

FEOM THE RED CRAG OF SUFFOLK. 217

curved, and with evidence of a large conical basal pulp-eavity. The specimen drawn in
fig. 7 is a somewhat large example referable to this group.

As parallel to the second group of recent Walrus-tusks, I am able to cite the beautiful
example drawn in Plate XXII. fig. 2, besides numerous fragments. The relatively large
curvature, the relatively slender proportions, the great lateral compression, and the
feebly marked surface-fluting, as compared with tusks referable to the next group, are
unmistakable characters in which this set of the fossil tusks agree with those known to
belong to females of the living Trichecus.

The largest, most massive, longest, most strongly ridged specimens which are found
in Suffolk, agree with the similarly characterized male tusks of living Walrus in
being also less curved than are those of the female. A fine example of the male tusk
from Suffolk is drawn in Plate XXII. fig. 1. The base of a still larger specimen, indi-
cating the largest Walrus-tusk, recent or fossil, which I have encountered *, is drawn in
fig. 3. In fig. 4 is drawn the point of a tusk, probably that of a male (on account of its
strong fluting), although exhibiting much greater lateral compression than is usual in
the tusks of the living male Walrus. Such flattened tusk-points are not uncommon in
Suffolk, and possibly indicate a Walrus-like animal distinct from that with less-com-
pressed tusks.

Lastly, we have from Suffolk specimens corresponding exactly to the 4th group of
tusks noted above as distinguishable in the collection of the Royal College of Surgeons.
One of these is drawn in fig. 5, and another in fig. G. They are distinguished by the
wide base, filled with a solid osteodentinal core and a (comparatively) very rapid tapering
of the tooth, which is relatively short and stumplike. These are undoubtedly old worn-
down tusks, and may be either those of males or of females.

Observations on the Curvature and Maximum She of recent and fossil Walrus-tusks, and
on the Fluting and Proportions as exhibited by Transverse Sections.

I obtained a number of outlines of the transverse sections of Walrus-tusks preserved
in the museum of the Royal College of Surgeons by means of modelling wax applied as
a ring to the tusk ; the ring was removed after pressure, and a pencil was drawn around
the inside of the ring when held on a piece of paper. These transverse sections may be
compared with those of the fossil tusks, and give the best means of estimating the differ-
ences which the two sets of tusks (the living and the fossil) present in respect of fluting
and transverse proportions.

It is to be noted that the fluting of the recent tusks is very varied in the different
specimens studied. The tusks are always oval in section, being somewhat compressed
laterally. One may find a single groove only on each side of the tusk ; or there may be
two grooves on the outside of the tusk and none on the inside, or there may be merely
the faintest indication of grooving (see figs. S to 14).

The same variation obtains in the fossil tusks from Suffolk ; some are more, some less
deeply fluted. But, as compared with the tusks of living Trichecus, the tusks of Tri-

* A specimen of the tusk of T. rosmarus recently acquired by the Royal College of Surgeons' Museum, to which
my attention has been called by Frof. Flower, is fully as large as this.

218 PROF. E. EAT LANKESTER ON WALRUS-TUSKS

cliecus Huxleyi show a somewhat deeper and more constant fluting ; whilst, further, it
is distinctly to be noticed that the transverse section is narrower than that of T. ros-
marus ; that is to say, the fossil tusks are more compressed laterally than are the living
ones.

It is not possible to give the curvature of the Walrus-tusks in a numerical form ; but,
as compared with those of living Walrus, the fossil tusks exhibit a more marked curva-
ture. The tusks of living Walrus vary in this respect very greatly. Whilst the tusks
of females are more curved than those of males as a rule, we also find that some male
tusks are nearly straight, whilst others are curved. As to absolute size, there can be no
doubt, from the measurements which I have taken, that the fossil tusks are larger than
the average of recent specimens. The largest tusk in the College-of-Surgeons' Museum
is 26J inches long *, as measured in a straight line, between the extreme points of base
and crown. The fossil specimen of the tusk of a male Trichecus Huxleyi drawn in fig. 1
would have measured, when complete, as much ; the fragment drawn in fig. 3 indicates
a much larger tusk ; whilst the tusk of a female drawn in fig. 2 is larger than any recent
female tusk which I have examined.

APPENDIX, July 17th.

The varieties of form presented by the tusks of the recent and fossil Walrus are due,
as pointed out above, among other things, to two independent conditions, which may
affect any two tusks compared in different degrees, viz. attrition and growth. Growth
does not necessarily proceed pari passu with attrition ; and consequently tusks of the
same age may be of various lengths, owing to the varying amount of attrition to which
they have been subjected.

The growth of the tusk consists in a continual addition to its base, that which was
the base or socketed portion of the tusk at one period of life becommg the naked and
even the terminal portion at a later period (see woodcut, where 1, 2, 3, 4 represent suc-
cessive bases with pulp-cavity). At a certain period of life the tusks cease to grow, as
is indicated by the reduction and even complete filling of the shallow pulp-cavity. The
biggest tusks will necessarily, ceteris paribus, be those which, having ceased to grow,
have during the period of growth suffered least from attrition.

The mode of wearing-down of the tusk is made obvious from the examination of worn
tusks, when the new outline formed by the wearing-down of the tusk is found to have a
certain relation to the fluting of the surface, which, in an unworn tusk, is parallel with
the unworn outline. It is found that the wearing of the tusk is greater on the convex
or anterior border of the tusk than it is on the concave or posterior border, and is so
considerable on the convex face as to preserve the tapering form of the tusk although
the point is rapidly worn away. This is shown in the woodcut, where the successive
outlines, A, B, C, D, indicate a series of stages in the attrition of the tusk. If the full-
grown tusk had never suffered in any way from attrition, it would present the complete

* The specimen referred to in a previous note, however, measures 31 inches in length along the chord, and 34 ,:
inches alona; the sreater curvature.

FEOM THE EED CEAG OP SUFFOLK.

219

outline shown in the figures. A combination of such base-lines as 2 and 3 with such
successive attrition-outlines as B, C, D, gives the possible variations of a tusk traceable
to the causes of " growth " and " attrition." The base-line 1, with the unworn point-
outline A, gives the form of very young tusks.

Diagram to illustrate the growth and attrition of the Tusk of Walrus: —
Fig. 1 with less, and fig. 1' with greater primitive curvature. Oscl, osteodentine ;
(hit, dentine ; A, original point of tusk; B, C, D, successive positions of
the point, resulting from attrition; 1. Early position of the base of tusk
and pulp-cavity ; 2, 3, 4. Successive positions of the base-line and pulp-
cavity due to new growth.

It will be observed that the successive attrition-outlines cut into the osteodentinal
core of the tusk ; and this fact enables us, by a careful inspection of the surface of a
Walrus-tusk, to determine approximately the amount of wear to which it has been sub-
jected. A short tusk is not necessarily an old tusk ; but if it have not its osteodentine
largely exposed on the convex border, its form is due to such a combination of
growth and attrition as we should obtain in the woodcut by taking the base-line 2,
together with the attrition-outline B.

When we add to the varied combination of various conditions of growth and attrition
the primary individual variations of the tusk, viz. greater or less curvature, greater or

220 PROF. E. RAY LANKESTER ON WALRUS-TUSKS

less lateral compression, greater or less amount of fluting, we have set forth the whole
series of elements which are factors in the form of any given Walrus-tusk.

The effect of a greater primary curvature on the appearance of the tusk at different
stages of growth and attrition will he obvious from the second woodcut, one remarkable
fact being that a short tapering tusk may be produced at an early stage of growth
and attrition, not unlike that which is the final stage (outline 4 D in fig. 1) in the
straighter tusk.

The causes of attrition of the Walrus-tusk are not found in the mere rubbing of the
points upon ice studded with sand particles, but in the longitudinal movement of the
tusks in digging up the sea-bottom when the Walrus is in search of Mollusca, and
very probably also in the longitudinal movement of the tusks against rock-surfaces when
they are employed to scrape off attached Mollusca, such as Limpets.

That the tusk-tissues are very easily worn dowu is proved by the condition of a skull
in the collection in my charge at University College, London, where the points of the
two tusks in a skull of Walrus have been worn flat by movement on the shelves and
tables of the Museum.

An important fact in this connexion is the nature of the sea-bottom and sea-shore
which were frequented by the fossil Walrus of the Crag. It is an established conclusion
that this late Miocene (Diestien) sea, with its Pyrula, Voluta, Cassidarla, Pholadomya,
and such forms, and its Teuthopbagous Whales (Zrpkioids) and its huge Sharks, was
not an ice-bound sea. The tusks of the Walrus, then, are only secondarily, and not
primarily related to its movements upon shore-ice.

At the same time, from what we know of the structure of this part of Eurojoe, it is
certain that the Diestien Walrus had no very hard rocks against which to wear down its
tusks. ^

And accordingly we find (as it seems to me) a less extensive attrition of the fossil than
of recent Walrus-tusks, and, as I have pointed out above, a greater primitive curvature
and a greater lateral compression. The combined effect of these special factors in the
case of the fossil tusks accounts for the differences of form which some of them present
when compared with tusks of the Living Walrus.

FROM THE RED CRAG OF SUFFOLK. 221

DESCRIPTION OF PLATE XXII.

Fig. 1. Tusk of male Trichecus Huxley i from the Suffolk bone-bed at the base of the Red Crag, nat.
size. The specimen is preserved in the Ipswich Museum.
1 a. Transverse section of same, taken at base.

1 b. Transverse section taken at the point.

Fig. 2. Tusk of female Trichecus Huxleyi from the Suffolk bone-bed at the base of the Red Crag, nat.
size. The specimen is preserved in the Ipswich Museum.

2 a. Transverse section at the base.

2 b. „ „ about basal third.

2 c. „ „ about distal third.
2d. ,, „ near the point.

Fig. 3. Tusk of a very large male of Trichecus Huxleyi from the Suffolk Crag, nat. size. From a cast in
the British Museum : the specimen was formerly in the collection of Mr. Baker of Wood-
bridge, and is now in the York Museum, Reed collection.

3 a. A transverse section near the base.

Fig. 4. Point of probably a male tusk from the Crag, nat. size, remarkable for great lateral compression.

Ipswich Museum.
Fig. 5. Old worn-down tusk from the Crag of Suffolk, nat. size. Ipswich Museum.

5 a & 5 b. Transverse sections of base and apex of the same specimen.

Fig. 6. Another worn-down tusk from the Suffolk Crag, also in the Ipswich Museum, nat. size.

6 a. A transverse section near its point.

Fig. 7. An incompletely grown tusk from the Suffolk Crag, nat. size. Ipswich Museum.

7 a. Transverse section at its proximal end.

Fig. 8. Transverse section of a male tusk of Trichecus rosmarus, 26A inches long, taken at 6 inches from

the base.
Fig. 9. Same tusk as fig. 8, transverse section at 4 inches from the point.
Fig. 10. Transverse section of a young tusk of Trichecus rosmarus, 8 inches long, taken at 2| inches

from the base.
Fig. 11. Transverse section of a male tusk of T. rosmarus, 21 inches long, taken at 5 inches from the base.
Fig. 12. Transverse section of an old worn tusk of Trichecus rosmarus, 10 inches long, and of conical

form, taken at li inch from the point.
Fig. 13. Similar section at 4 inches from the point.
Fig. 14. Similar section at 6 inches from the point.

SECOND SERIES. — ZOOLOGY, VOL. II. 30

[ 223 ]

VII. The Parasites of Elephants. RyT. Spencer Cobbold, M.J)., F.R.S., F.L.S.,

Foreign Member of the Royal Agricultural Academy of Turin.

(Plates XXIII. & XXIV.)

Bead April 7th, 1881.

Preliminary Remarks.

CONSIDERING the importance of Elephants as domesticated animals and the
abundant opportunities for post-mortem examinations that occur in India, it is
surprising how little has been done towards completing our knowledge of the parasites
of these large quadrupeds. As we proceed it will be found that most of the facts of
parasitism at present known have been gathered from the examination of animals
that have died in captivity either in Europe or in America.

Eor some years past I have been engaged in collecting materials, which thus far have
only served to supply occasional and very incomplete records ; moreover the notices in
question have chiefly appeared in a periodical inaccessible to most naturalists.

Either from the mistaken notion that the study of parasites is unattractive, or that it
is unremunerative, this subject has been strangely neglected ; for there is no animal of
equal importance with the Elephant about whose parasites so little is known. Naturalists
abroad have done almost nothing in this direction ; and of those at home one noteworthy
exception alone occurs in the person of the late Dr. Baird, who described several species
contributed by Dr. Murie and Mr. Gerrard respectively to the British Museum. The
entozoa preserved in that institution, however, have been for many years past practically
inaccessible to investigators.

It is now desired to supply a summary of all that has hitherto been published, cor-
rected and extended by data that are entirely new. Only such anatomical details as
present points of zoological interest can be given ; but it is proposed to supplement the
natural-history facts with certain particulars which serve to explain the r61e of parasites
in the production of fatal epizootics. This method, with the Society's approval, may
stimulate further inquiry, and thus ultimately lead to the adoption of measures cal-
culated to lessen the mortality amongst domesticated Elephants.

The materials placed at my disposal for this investigation have been contributed at
different times, and severally, by Professor Huxley, by Assistant Commissary-General
Colonel H. P. Hawkes, of the Madras Staff Corps — whose earlier contributions reached me
through the late Dr. Baird and Dr. Hugh Cleghorn. — by Veterinary Surgeon Thacker,
also of the Madras Army, by Mr. John Henry Steel, A.V.D., late Demonstrator of
Anatomy the Royal Veterinary College, and by Veterinary Surgeon Frederick Smith,
who had professional charge of some Elephants that died at the stables attached to

SECOND SERIES. — ZOOLOGY, VOL. II. 31

224

DR. T. S. COBBOLD ON THE PAEASITES OF ELEPHANTS.

Sanger's Circus. An independent examination of the viscera of one of the Circus
animals likewise supplied me with several new species of internal parasites. I am also
indebted to Prof. Flower and Dr. Garson for permitting the examination of two
tusks preserved in the Hunterian Museum.

At the time that Dr. Baird wrote his first paper (1859) only two species of Entozoa
were supposed to infest the Elephant ; and one of these had been imperfectly described.
He then added a third. In 18G8 Dr. Baird indicated a fourth species ; and one year
later the hitherto imperfectly but best-known parasite was pretty fully described by myself.
Thus up to the year 18G9 only four internal parasites were known to infest Elephants ;
and of these one had been referred to the genus Ascaris, two to the genus Sclerostoma,
whilst the fourth species was supposed to be identical with the common fluke. This was
an error. To show how little progress has since been made, I may observe that when,
so lately as 1878, Dr. O. von Linstow published his useful ' Compendium der nelmin-
thologie,' the actual number of known and really genuine species had not increased. To
be sure, Dr. von Linstow did not hesitate to record the occurrence of six distinct species ;
but he accomplished this result by placing the fluke parasite under three separate specific
names. There is not the smallest ground on which to justify this process of species-
splitting. If my determinations are correct, it follows that up to the present time only
four Entozoa proper to Elephants are known to science ; however, notwithstanding my
reduction of Dr. von Linstow's list, and the consequent rejection of several recorded
species, I am now enabled to furnish proofs as to the occurrence of at least twelve
distinct species of internal parasites in the Indian Elephant. This includes the bots or
larva3 of a dipterous insect . Of the external parasites I have made no special study ; but
if these be added to the parasitic fauna, the total number of Ectozoa and Entozoa com-
prises not less than fourteen good species. A simple table will best express their
zoological position and relations.

Entozoa

m

Eh

i— i

<

Ectozoa

Ascaris lonchoptera, Diesing

Sclerostoma sipunculiforme, Baird..

Strongylus clathratus, Baird

Strongylus foliatus, sp. nov

Strongylus falcifer, sp. nov

Dochmius Sangeri, sp. nov

Filaria Smitkii. sp. nov

Amphistoma Hawkesii, Cobb

Amphistoma ornatum, sp. nov

Amphistoma papillatum, sp. nov. ..

Fasciola Jaclcsoni, Cobb

Gastrophilus elephantis, Cobb

Ova of a dipterous insect (Sclater)
Hcematomyzus elephantis, Piaget ..

Nematoda

} IIELMINTHA.

TllEMATODA ,

Insecta

ARTHROPODA.

Homopus elephantis, Meguin A rachnida ,

Eor reasons that will be given later on, I have not reckoned Sclatcr's dipterous ovum
as representing a separate species.

DE. T. S. COBBOLD ON THE PAKASITES OF ELEPHANTS. 225

Description of Species, and Comments.

1. ASCAEIS LONCHOPTERA, Diesing.

A.lonchoptera, "Diesing, Syst. Helm. Bd. ii. S. 176, Snpp. S. 560; idem, Denkschrift. d. math.-nat. CI.

d. k. Akad. d. Wissenschaften, Bd. xiii. S. 13 (with figures).
Ascarides, Jackson, Catalogue of the Anat. Mus. &c., Boston, 1817, p. 317.
Strongylus elephanti (sic), Rudolphi, Synopsis Entozoorum, Berlin, 1819, p. 36.

Body of uniform thickness, but rather narrowed in front and behind, with an acutely
pointed conical tail. Head with three large almost spherical lips, the margins of the
neck having two lateral semilanceolate winged appendages.

Male unknown.

Length of the female 3 inches (or more). Breadth fully £ of a line.

Sab. Biliary ducts and duodenum of Mephas indicus.

The only published description of this worm is that given by Diesing. He states that
the parasite was discovered quite accidentally by Herr Mayor in the year 1822. Several
parasites (but it is not stated how many) were found in the gall-ducts of an Asiatic
Elephant. This animal had been held in captivity at Geneva, where it was destroyed on
account of madness (wegen Lobsucht). Not until the year 18-17 do we find any second
record of the occurrence of this entozoon. In this instance several specimens were
obtained from the biliary ducts and from the duodenum of an Indian Elephant which
died in captivity at Boston, TJ. S. A. It is not stated how many of the parasites existed ;
but Dr. J. B. S. Jackson, to whom both the find and its record are apparently due, states
that these Ascarides were associated with numerous flukes. As Dr. Fitz has recently
pointed out, the flukes themselves were at the time considered to be examples of the
common liver-fluke of ruminants (Disiomum hepaticum).

It would be interesting to have further particulars respecting the American " find."
Of the Geneva specimens, only two, both females, were sent to Vienna ; and these were
originally entered in the Catalogue as Strongyles. As at present we possess no published
description of the male parasite, it may be that the Boston-Museum specimens would, if
carefully examined, supply us with the missing link. Even the few particulars that we
have respecting the characters of the female worm require verification. The specimens
noticed by the illustrious Rudolphi were the same as those described by Diesing ; and
supposing that Iludolphi actually examined the worms, it is difficult to understand why
he should have called them Strongyles, or rather, why he should have placed them
amongst his species dubice, and, as such, in the genus Strongylus. The conspicuously
three-lipped mouth ought to have afforded sufficient proof of its true generic position.
But for Diesing's figures, one might still have remained sceptical as to the ascarid
character of the Geneva specimens ; and certainly as regards the nature of the Boston
find I am still somewhat in doubt, all the more so since the specific characters of the
flukes found with the Ascarides were from the outset misunderstood. Perhaps Dr. Eitz,
who has cleared up certain points respecting the organization of the flukes, will set this
matter at rest. At all events, until our own naturalists in India bestir themselves and

31*

226' DE. T. S. COBBOLD OX THE PAEASITES OF ELEPHANTS.

make fresh finds, it may be said that the first helminth discovered to infest the Elephant
is just that very species about whose structure and economy we are least informed.

2. Scleeostoma siPUNCTJLiFORME, Baird. (Plate XXIII. figs. 1, 2, & 0, 7, 8.)

S. sipunculiforme, Baird, Proceed. Zool Soc. 1859, p. 427; also P. Z. S. 1868, p. 262 (with figures).

Body smooth, with extremely fine transverse striae, sipunculiform, thickset. Head
distinct, separated from the body by a distinct neck or constriction. Mouth circular,
armed with numerous upright denticles arranged in two rows, with four larger teeth, one
on either side, and one in front and behind, springing from between the denticle-rings.
Bursa longer than broad, apparently supporting twenty rays, the anterior ray being
doable, the posterior ray quadruple, its innermost division supporting two rudimentary
branches of unequal size; posterior lateral ray short. Spicules equal, long, straight,
and narrow. Tail of the female obliquely truncate, the extremity conical and suddenly
tuxmed backwards. Reproductive outlet immediately above a projecting lobe or tubercle
in front of the tail. Anus situated in a deep hollow at the base of the cone-shaped end
of the tail.

Length of male nearly 1 inch (11 lines).
Length of female above 1 inch (14 lines).

Hah. Intestines of JElephas indicus.

If this description be compared with that given by Dr. Baird, it will be seen that our
records differ in several important particulars. As regards the position of the repro-
ductive outlet of the female, I have relied upon his authority ; but in respect of other
characters, especially those of a microscopic kind, my interpretations of the facts observed
are wholly different.

As the study of the helminth-fauna of Elephants is in its infancy, it is necessary to be
precise in the few details at my command. According to Baird the body is " cylindrical
in shape, thicker in the middle, tapering towards each extremity, and finely striated
across, though the stria; are rather distant from each other." This is correct, except
that Baird's description of the distant stria; must, I think, refer to two or more rings in
the neighbourhood of the neck, since the ordinary stria; placed between these rings and
all over the body are excessively delicate markings. They cannot be clearly seen under
low magnifying powers ; and therefore I have not attempted to represent them in my
enlarged representation of the neck and head. In respect of the head itself, Baird says
it is " rather large, cylindrical, about one line in length, thicker than the neck, and
separated from it by a distinct line or groove. This groove I find to be succeeded by
other similar rings below it ; and as they are tolerably distant from each other, I think
Dr. Baird must have referred to them when he spoke of the stria? as being rather widely
separated. The remainder of his description of the head stands as follows : — " The
mouth is orbicular, placed in the centre of the truncated part of the head, and
surrounded with two homy capsules or bullae, the limbs or mai'gins of which are each
armed with a row of numerous teeth. The limb of the external capsule is the larger of
the two, and the teeth, though numerous, are less so than in the internal limb, and are
stronger and pointed outwards. The limb of the inner capsule is much smaller, the teeth

DP. T. S. COBBOLD ON THE PARASITES OF ELEPHANTS. ^27

very minute, and exceedingly numerous. This portion of the body does not differ in the
two sexes."

With this part of Dr. Baird's description my observations generally agree ; but he
appears to have overlooked the four larger equidistant denticles, of which the lateral pair
are especially conspicuous. I examined the mouth with some care. By a little pressure
the buccal apparatus and narrow muscular oesophagus are well seen under a half-inch
objective, several abductor and adductor muscles being plainly visible through the
transparent skin. The buccal cavity itself is broadly cup-shaped, being armed, as
Baird has said, with two rows of closely set and erect denticles. I did not observe any
marked disparity of size, as between the two sets of teeth ; and their number is probably
the same. I twice counted the denticles of the outer circle, and found their number to
be 44 ; consequently, with the four larger teeth, the mouth cannot be armed with less
than 92 of these organs. The length of each tooth was about the -^s °f an mcu °nly,
the teeth of the inner circle alternating in position with those of the outer.

The bursa, or hood, of the male worm is a very striking object. By reason of the
unusual number of rays or, rather, I should say, of ray-divisions, a correct interpretation
of these parts is not easy. Dr. Baird's account and that which I am about to offer are
scarcely reconcilable. If my account should prove the more correct, it must be remem-
bered that Dr. Baird had not the advantage of knowing Schneider's published views
respecting the typical number of parts to be recognized in this organ. It is difficult to
interpret the facts given in Dr. Baird's memoir, though they have evidently been
recorded with great care.

" The tail of the male," says Dr. Baird, " is in the form of a membranous expansion
or pouch surrounding this extremity of the body, and is divided into three lobes. The
central lobe is the largest, and is supported by seven ribs or rays, three in the middle
and two at each side. The three central ribs are dichotomous, and the middle one of
these has its two branches sending off two or three short processes, like buds ; the two
others are simple. Of the lateral ribs (two on each side) the innermost one is dicho-
tomous, the two branches into which it is divided sending off short processes or buds ;
the outer rib is simple. The two lateral lobes of this caudal expansion are smaller thau
the central one, and are each supported by four simple ribs."

In the right and left views (profile and oblique respectively) which I have selected for
general illustration of the appearances presented by the tad of the male worm, little
more than the points of the rays can be seen (PI. XXIII. figs. 2 & 8). As there
shown, nearly all the rays extend to the border of the hood, whilst some of them
point beyond the ordinary level of the margin, which consequently assumes a slightly
serrated outline. This appearance is not constant. Having only two specimens of the
male to examine, I found some difficulty in arriving at a true interpretation of the parts ;
but I believe they are correctly rendered in the accompanying scheme. It is, of course,
immaterial to which half of the hood-rays the description is taken to refer, since, prac-
tically, the arrangement and symmetry are perfect throughout. The anterior ray (a, a')
is dichotomous and cleft to the base. The antero-lateral ray (b) is straight and simple.
The middle ray (c, d) is dichotomous, cleft to the base, andthe branches incline toward

228

DE. T. S. COBBOLD ON THE PAEASITES OF ELEPHANTS.

Fig. 1.

Diagram of the hood and rays of Sclerostoma sipwicidi-
formt ( x 36 diameters i.

each other at their points. The postero-lateral ray (d) is simple, and rather short. The
posterior ray is furnished with a remarkably broad base (e) ; it is tetrachotomous
(e to e""), its innermost branch being
united at the upper part to its fellow of
the opposite side. The innermost branch
gives off two rudimentary secondary
branches, the inner of the two necessarily
lying nearly parallel to that of the opposite
side, and therefore also parallel to the
intervening middle line forming the ver-
tical axis of the hood. As regards the
membrane of the bursa itself, the three
lobes are fairly distinct, the posterior
lobe being large, well marked, and slightly
truncated at the apex.

If the above description be compared
with that given by Dr. Baird, a corre-
spondency may be traced here and there.
Thus his so-called middle ray of the cen-
tral lobe evidently agrees with the inner
division of the posterior ray in Schneider's

system (as applied to the present species). Thus, also, the two or three bud-like processes
to which Baird refers must clearly be the same as those rudimentary rays observed by
myself, and called the outer and inner secondary branches of the innermost division of the
great posterior ray. Of course in attempting to reduce and describe this complicated set
of rays to the proper number assumed by Schneider to be typical of the strongyles, my
interpretation may be open to dispute. I do not think there is any important error in
this record, although the occurrence of such a number of separate ray-divisions in
any single hood is altogether outside my previous experience.

The spicules of Sclerostoma sipunculiforme are long, equal, and remarkably narrow.
They measure about ^ of an inch in length, and have their points irregularly twisted, and
directed backwards. In their retracted state they are apt to be overlooked, although gentle
pressure of the covering glass will readily reveal their presence and position (PI. XXIII.
fig. 2). When partly exserted, their shafts protrude from a cloacal outlet immediately
in front of which there is a conspicuous prominence (PI. XXIII. fig. 8). Neither in Dr.
Baird's figures nor in his description of this worm is there any mention made of the
existence of the spicules. Doubtless they were retracted within their sheaths, and
thus escaped observation. As these organs vary remarkably in the different species
of strongyles infesting Elephants and other animals, no diagnosis can be considered
complete if all mention of their characters be omitted.

Dr. Baird's description of the tail of the female worm agrees generally with what I
have observed. " It terminates," as he says, " in a rather long sharp point, which is
oblique in position to the body, owing to a sort of tubercle about half a line from its

DE. T. S. COBBOLD OX THE PAEASITES OF ELEPHANTS. 229

extremity, under which is situated the anus. This aperture is very distinctly seen
immediately underneath the tubercle ; and tbe vulva, which is not very conspicuous,
and is in the form of a narrow slit across the body, is situated immediately above the
tubercle."

Seen in profile from the side, the observer cannot fail to remark upon the resemblance
which the tail of the female shows to the outline of a Chinese lady's foot ; the narrow
lower end of the body having the form of the ankle, and the so-called tubercle corre-
sponding with the heel (PL XXIII. figs. 6, 7). I made no special examination of the
internal anatomy of this species ; but so far as the transparency of the body permitted,
the arrangements seen by me corresponded with what Baird has stated to occur. " The
oesophagus is rather long, and terminates in a lobed stomach which extends for a short
distance and then ends in a straight intestinal canal running through the whole length
of the animal. The uterus is peculiar in form, and presents a very pretty appearance
under the microscope ; it is two-branched, and has during its length several expansions
or swellings followed by contractions. The ovaries are very long and twisted round the
intestine."

In the course of his paper Dr. Baird calls special attention to the relative size of the
sexes, more particularly with the view of comparing the characters of this species with
those displayed by certain well-known Sclerostomata from the horse (S. armatum and
& tetracanthus) and other animals. In Baird' s specimens the males were longer than
tbe females — a circumstance which, as his remarks imply, is of most unusual occurrence.
There may have been some mistake here ; at all events, in my set of specimens the
females are longer than the males. Not improbably Dr. Baird's female worms were
immature. Another curious feature connected with the external anatomy of this species
related to the presence of certain markings beneath the skin. These appearances were
evidently seen by Dr. Baird, who, strange to say, refers them to the species which I shall
next descrihe. More than this, they are actually figured in one of his otherwise gene-
rally accurate woodcuts, and are held to be of value in reference to the identity of the
species. Thus, as will be seen, Dr. Baird has selected this very striking appearance as
affording sufficient grounds for the special nomenclature which he proposed for the
following species. Accordingly he has named that worm Sclerostoma clathratum, on
account of the clathrate or lattice-like markings which may be seen at or near the
surface of the body. Now, after a long study of this matter, I am obliged to conclude
that the enlarged figure of the head of the female worm given by Baird, in which the
appearances in question are represented by no means accurately, is not that of his
second new species, but really and truly a representation of the head of the species now
under consideration. If this surmise be correct, it might seem necessary to alter Baird's
nomenclature ; but although, in my view, the specific name applied by him is altogether
a misnomer, I propose to retain the nomenclature which, by some oversight, he con-
cluded to adopt.

The appearances in question are very peculiar. They are not clue to external sculp-
turing, as Baird's figure would imply, but to the presence of an internal network,
limited, as I think, to the surfaces of the longitudinal muscles. Under low magnifying

230 DE. T. S. COBBOLD ON THE PAEASITES OF ELEPHANTS.

lenses the markings pretty closely correspond with the polyhedral networks figured as
well as described by Schneider from two species of Gordius. In our Elephant's
nematode the network seems to lie somewhat deeper (PL XXIII. figs. 1, 2). To myself
the finer markings present the appearance of vascular ramifications proceeding from
common stems, each muscle having a group of its own. There are thus several sets of
vessels; and these groups, as such, do not anastomose with each other. Probably there
are main trunks connecting all the secondary stems together ; and, taken as a whole,
they have not the uniformly reticulated characters seen in the network of the Gordii.
I believe them to form special developments of the water-vascular system. Their
structure deserves further study. I should not, indeed, have hazarded thus much con-
cerning them, had not Dr. Baird emphasized their importance as a basis of specific
nomenclature. As it appears to me, either Dr. Baird or his artist has made the head and
neck of a female Sclerostoma sipunculiforme do duty for the next equally well marked
species; at least I find the clathrate appearance in this worm, and not in the worm
which he has called Sclerostoma clathratum. The two worms are undoubtedly good
species ; and their respective heads bear no sort of resemblance to each other, either as
regards form, size, markings, or oral armature. If Dr. Baird's own admirably clear
diagnosis be examined, one can hardly fail to see that his description of the head and
neck of Sclerostoma clathratum only corresponds with some of the particulars given in
the illustration to which I have referred. In making this statement I am far from
supposing that my own observations are free from error.

3. Strongylus clatiikattjs, Baird. (Plate XXIII. figs. 3, 4, 5.)
Sclerostoma clathratum, Baird, Proceed. Zool. Soc. 1868, p. 202 (with figs.).

Body linear, sub filiform, cylindrical, but rather suddenly narrowed in front and
behind. Head small, scarcely half the width of the body, truncated, surrounded by
five very small auricular folds. Mouth armed with a few teeth. Bursa broader than
long, obscurely three-lobed ; anterior ray short and double ; antero-lateral ray simple
and straight ; middle ray double, of very great size ; postero-lateral ray moderately
long and slender ; posterior ray cleft at the extremity. Spicules two, broad, and highly
coloured. Tail of the female bluntly pointed, with the reproductive outlet slightly
removed from the base in front.

Length of the male 2 inches.

Length of the female 2 inches to 3 inches.

ILab. Stomach and intestines of Elephas indicus and Loxodonta africana.

In my account of the previous species the origin of the nomenclature adopted by
Dr. Baird is explained ; and, excepting that I prefer to place this worm with the
strongyles, properly so called, rather than in the genus Sclerostoma, no alteration of the
title seems necessary. If our diagnosis be compared, it will be seen that whilst much has
here been added, little, save as regards the clathrate character, has been altered in
respect of Dr. Baird's description. With the exception of the enlarged head of the female
worm, all the figures given by Dr. Baud arc thoroughly characteristic, and probably

DE, T. S. COBBOLD ON THE PAEASITES OF ELEPHANTS. 231

accurate. That marked by him 2 a (woodcut P. Z. S.) is perplexing. The head is not
only altogether out of all proportion, but the slight cutaneous folds, which form mere
sculpturings or elevated lines, have in his figure all the appearance of large auriculate
lappets (see Plate XXIII. fig. 3). As already explained, lattice-like markings have not
been seen by me in this worm. Dr. Baird's life-size representations correctly show that
the head and neck are finely drawn out to a blunt point ; but the enlarged figure of the
head above referred to conveys the contrary impression. Although not expressly so stated,
his figures imply that the heads of the two sexes closely resemble each other in respect
of shape, if not in size. This is usually the case in Nematodes. Unfortunately I do not
possess a female Strongylus clathratus, and have therefore been guided in my inferences
from the characters presented in the head of the male worm. It follows from this that
what has here been said respecting the form of the tail of the female rests entirely on
Dr. Baird's authority ; and here also I may remark that his enlarged view of the tail of
the female, which I have no doubt is correct, presents no trace of the remarkable latticed
structure that forms so striking a feature in his enlarged representation of the head and
neck of the female. In the male Sclerostoma sipuncirfiforme this network occurs both
near the head and tail, and throughout the length of the body. It occurs in the female
iikewise ; but I have not shown it in my less highly magnified illustrations of the tail of
that species.

In reference to the habitat of this worm, Dr. Baird states that his specimens were
obtained by Dr. Murie from the stomach of a young female African Elephant {Loxodonta
qfricana). These Dr. Murie sent to the British Museum. The Elephant died in London
in October 1SG7 ; and Dr. Baird concludes his account of the worm by saying, " I am
not aware of any Entozoa having been previously described as inhabiting the body of
this species of Elephant." Dr. Baird's conjecture was perfectly correct ; and although
I have to record the occurrence of a dipterous larva in the African Elephant, yet even
up to the present time no other true helminth than that discovered by Dr. Murie has
yet been described from this animal. My own specimens of Strongylus clathratus were
obtained from the Indian Elephant (Elephas indicus), and were, I believe, all removed
from the small intestines. In Murie's find " the specimens were rather numerous, and,
with only one exception, all were females." It is stated that the male worm was white,
whilst the others were of a " dark brownish red colour, the contrast between the two
sexes being for a long time distinctly visible." As Baird says, on comparing this
parasite with the previous species, the head of Strongylus clathratus " is much smaller,
both in male and female, and the caudal bursa differs in structure from that of
Sclerostoma sipunculiforme." As supplementing the diagnosis, he adds the following par-
ticulars : — " The head is distinctly separated from the rest of the body, and is cylindrical
and small ; the mouth is dentated inwardly, and on the outer edge is marked with five
distinct auricles, which descend upon the upper portion of the body, but do not form a
distinct part of it. The body is attenuated at both extremities, and is linear in shape ;
in the female it is, as it were, minutely clathrate, being situated transversely with fine
lines, and having the spaces between the lines minutely longitudinally striated or
sulcated, giving, under tolerably high power, the appearance of being latticed or

SECOND SERIES. — ZOOLOGT, VOL. II. 32

232

DR. T. S. COBBOLD ON THE PARASITES OF ELEPHANTS.

clathrate. The tail is obtuse, not so sharp-pointed as in Sclerostoma sipunculiforme, and
has the genital organs situate a short distance from the extremity. It is frequently,
but not in all cases, convolute. The body of the male is longitudinally striated ; and the
bursa at the extremity of the tail is lobed."

Notwithstanding what I have advanced respecting the clathrate appearance, the fact
of my not having examined a female worm of this species will probably leave the question
still an open one. As regards the male, the characters of the hood are particularly
striking, even more so than they are in Sclerostoma svpu/nculiforme.

Whilst Dr. Baird's enlarged view of the bursa makes this sufficiently apparent, he has
not supplied any description of details. In the hood before us we find two enormously
exaggerated divisions of the middle
ray (c, c') ; but in Sclerostoma sipun-
culiforme we found that most of
the rays were of more or less uni-
form length and thickness. Here
also we have the anterior ray
double (a), with its halves closely
applied, as obtains in most Stron-
gyles ; we have the antero-lateral
ray (b) simple and widely sepa-
rated from those above and below
it ; we have the ends of the huge
finger-like divisions of the middle
ray (c, c) extending far beyond the
terminal points of the posterior
rays ; and we have both postero-lateral (d) and posterior rays (e) reduced to within ex-
tremely narrow limits. Any one at all familiar with the typical forms of Strongyle bursse
cannot fail to see at a glance that the morphological departures here displayed are not
merely distinctive of the species as such, but altogether unique in themselves. In this
connexion another noteworthy feature relates to the form of the spicules. Neither in this
nor in the preceding species do these organs appear to have come under Dr. Baird's ob-
servation. Whilst retracted they are not readily visible under the microscope ; however,
a strong light, aided by slight pressure, is sufficient to expose them to view. In this
species each spicule consists of a shaft which is truncate above and pointed below. The
lateral blade-like expansions are unequally developed, the right side being flabelliform,
diaphanous, and marked by branching stria?, which proceed from the axis and arc
radiantly disposed (PL XXIII. fig. 5). The spicules are of equal size, each measuring
Yq of an inch in its long diameter. Immediately above the hood, and in front of the
spicule, there is a slight enlargement, which, even in spirit-specimens, preserves a
yellowish colour (PI. XXIII. fig. 4). This is probably an unusually large seminal
reservoir.

Concerning the special part played by this and the preceding species in the production
of injury to the host, evidences will be produced at the close of this memoir.

Diagram of the hood and rays of Strongylus claihratus

( x 24 diam.).

DE. T. S. COBBOLD ON THE PAEASITES OF ELEPHANTS. 233

4. Strongylus foliattts, sp. nov. (Plate XXIII. figs. 9-13.)

Body smooth, of nearly uniform thickness, finely striated transversely. Head
indistinct, truncated in front, with five small auriculate folds. Month slightly ohlique,
leading to a deep buccal cup, armed with a few coarse teeth, succeeded by a long muscular
oesophagus. Neck marked by two short conical papillae at the upper part, one on either
side, and by two larger and longer tapering papillae placed lower down. Hood bilobed,
foliate. Rays few in number, and widely separated. Spicules long, flattened, and
twisted at the base, tapering and finely pointed below at the apes. Tail of the female
well marked, long and conical, with a wavy contour, directed backwards, and rapidly
narrowing to form a subulate point. Anus in front, and immediately above the base of
the tail.

Length of the male h an inch.

Length of the female f of an inch.

Sab. Tumours within the coats of the stomach of Elephas indicus.

On the 24th of August 1876, I examined part of the carcass of an Elephant at the
Royal Veterinary College. The viscera had been removed from one of several animals
that had died at the Circus stables of Mr. Sanger. Mr. Steel, during his previous dis-
sections of this animal, had already encountered some parasites ; but in addition to these,
I obtained others occupying tumours in the walls of the stomach. The morbid swellings
were of two sizes, the larger and more rounded kind containing the species under con-
sideration, the smaller and comparatively flat tumours being occupied by worms of a
different generic type.

On the external or mucous surface of one of the larger growths there was an opening
of the diameter of a crow-quill. This appeared to be the only communication between
the cavity of the stomach and the interior of the tumour. On section the growth dis-
played a coarsely alveolar structure, its freely intercommunicating loculi being filled
with thick mucus, in which a few parasites were found imbedded.

The swellings in question are very similar to the well-known verminous tumours,
occupied by maw-worms, of the horse ; the parasites, however, are perfectly distinct.
Through the help of Mr. Spooner Hart of Calcutta, I have had opportunity to examine
specimens of the growths removed from a horse that died at Calcutta stables. In Mr.
Hart's experience these formations are rather common in India ; but it would probably
be incorrect to assume that they are more prevalent in eastern countries than in Europe.
Be that as it may, it is scarcely more than forty years since M. Valenciennes found these
singular worm-tumours to exist in eleven horses out of the twenty-four animals whose
stomachs he specially submitted to examination. Not improbably, therefore, if medical
men, army veterinary surgeons, and others practically interested would make use of
their opportunities, the statistics of gastric parasitism, both as affecting elephants and
horses, would afford much curious instruction. The natural-history value of the facts
would lie further increased by similar investigations amongst the Rhinoceroses and other
large herbivora. Although Schneider and Molin have both described parasites taken from
similar tumours in the American Tapir, we as yet know absolutely nothing of the
entozoa infesting the Tapir of the Malayan peninsula.

32*

234

DE. T. S. COBBOLD OX THE PARASITES OF ELEPHANTS.

Fig. 3.

In reference to the reproductive apparatus of Strongylus foliatus, I am unable to state
the position of the female outlet, whilst the rather small hood of the male presents
appearances that are difficult of interpretation. The anterior ray (a) is clearly dicho-
tomous and normal in other respects. The next ray, or pair of rays (b), forms a puzzle.
Either the antero-lateral ray is altogether suppressed, or, if present, it is dichotomous
and very similar in appearance to the anterior rays, the two halves being closely applied.
According to Schneider's mode of interpretation, the antero-lateral ray is invariably
single ; and thus, if his view be correct, the lower half of the ray (c) must, in the parti-
cular case before us, belong to the middle ray. Strange as it may seem, this is probably
the correct interpretation of the facts here exhibited.
From this it will follow that the next two large and
conspicuous rays (c1, cl) are not (as one might have sup-
posed) the divided shafts of a great dichotomous middle
ray, but the upper division (c') represents the lower
segment of the middle ray (as in Schneider's system),
and the lower division will thus represent the usually
solitary postero-lateral ray (d). The latter ray, accord-
ing to Schneider, is also always single. In this fashion
I am contented to interpret the facts, though, looking
to the exaggerated ray-divisions in question, it is diffi-
cult to disassociate them so as to conform to Schnei-
der's plan. It seems as if the large pair of half-rays

must correspond writh the yet more exaggerated half-rays of the great middle ray of
Strongylus clalhratus. When we come to the posterior ray (e) all difficulties of homology
vanish. In Strongylus foliatus it is dichotomous or deeply cleft, the base being united
to its fellow of the opposite side.

The spicules, though offering nothing remarkable in structure, arc characteristic of
the species. At the base they are bent in a spiral manner, and slightly flattened. Their
shafts lie parallel to one another, and become narrowed very gradually towards the tips,
where they are slightly turned forwards and finely pointed. Each measures about ^s °f
an inch from end to end.

In regard to the membrane of the hood, I should have stated that under moderately
high powers one may observe regularly arranged and excessively delicate striye radiating
from the base to its circumferential border.

Diagram of the hood and rays of
Strongylus foliatus ( x 60 diain.).

5. Strongylus falclfer, sp. nov. (Plate XXIII. fig. 11, and Plate XXIV. figs. 1-3.)

Body of a bright orange-red colour anteriorly, more or less curved, smooth, and rather
narrowed in front and behind. Head well marked, short, with radiating external lines,
and truncated. Mouth simple, surrounded by an outer series of four large conical
papillae and by an inner circle of numerous small papilla? or granulations ; oral cavity
shallow. Bursa conspicuously three-lobed, with the posterior rays large and dichoto-
mous. Sj/icules long, strongly bent towards their ends, which are blunt, spatulate, and

DE. T. S. COBBOLD OX THE PAEASITES OF ELEPHANTS.

235

Fig. 4.

directed backwards. Tail of the female very distinct, scythe-shaped, and finely pointed.
Reproductive outlet placed at a short distance above the anus, at the base of the tail.
Length of the male 1 inch (or 11 to 13 lines).
Length of the female 1 to If inch (or 12 to 20 lines).

Hob. Intestines of JSlephas iiul/cus.

With the unassisted eye this species is readily recognized by its bright colour, the
orange-red tint about the head gradually becoming feeble towards the anterior third
of the body. Out of two dozen examples of this worm I only secured two males ; and
these were picked out from amongst a quantity of other nematode species. The worms
had not been well preserved ; and their bodies had become coated with fungi.

Not wishing to mutilate the only two male worms I possessed, my interpretation of
tbe appearances of adhesion and union at the base of the bursa may be partly incorrect.
So far as I made out, the rays not only conform to the type-number, but present little
that is striking either as regards their shape or their arrangements. Speaking generally,
those of the side lobes of the hood are short and stumpy, whilst those of the middle lobe
are compararatively large and long. The anterior ray (a) is dichotomous, with adherent
divisions. The antero-lateral ray (b) is simple
and normal in position. The middle ray (c c') is
bifurcated, and has its branches rather widely
apart. The postero-lateral ray (d) is compara-
tively long and well within the middle hood-lobe.
The posterior ray (e e') is unusually large, and
divided into subequal halves, both of which nearly
reach the lower border of the hood. A consider-
able space separates the inner branch from its
fellow of the opposite side.

The spicules form a striking object. Viewed
from before or in front, the two shafts present a
lyrate figure, owing to their lower ends being
bent first outwards laterally, then inwards late-
rally, and finally, in an oblique direction, backwards divergently (PL XXIV. fig. 2).
Seen in profile the shafts appear rather narrower ; and they appear to lie parallel to each
other, except at their somewhat flattened and blunt points. The sheaths in which the
spicules usually lie, more or less concealed, can be seen within the half-transparent body,
powerful adductor muscles being inserted at the upper or closed ends of these pouches.

The body of the male at the lower part measures about the -^ of an inch in transverse
diameter, whilst the spicules, which are of equal size, afforded a length of fully ys °f an incu-

The best-preserved female worms had their oviducts crowded with eggs. Each
ordinary ovum gave an average long diameter of -g^y by -g-Jo of an inch in breadth.
Such eggs contained a dense and finely segmented yolk. On one occasion I observed
a solitary empty egg-envelope, which gave a measurement of no less than t\q of an inch
from pole to pole, by 2^u of an inch in breadth. At the same time, and certainly from
the same source, the field of the microscope displayed an embryo that had evidently escaped

Diagram of the hood and rays of Strongylus
faldfer ( x 43 diaiu.).

236 DE. T. S. COBBOLD ON THE PARASITES OF ELEPHANTS.

from this ruptured envelope. The embryo measured -7\- of an inch long by something less
than -5^0 of an inch in breadth. Although there were no intermediate stages observed
between the two kinds of ovum, I could not regard the presence of the much longer egg
and its embryo as merely resulting from accident. Before I examined these worms they
had been carefully separated from the other nematodes. As regards the uterus itself, I
have only to add that the two horns may be easily traced downwards to their junctions
together and with the oviducal or vaginal passage (PI. XXIV. fig. 3). Distinct
openings, with very slightly elevated margins, mark alike the position of the vulva and
the anal aperture.

6. Dochmitjs Sangeri, sp. nov. (Plate XXIV. figs. 4-0.)

Body smooth, transversely striated, and narrowed both in front and behind. Head
very distinct, and bent well forward. Mouth simple, oval, and surrounded by regularly
disposed elevations; buccal cup capacious, and armed with a few small teeth; pharynx
distinct, separated from the oesophagus by a well-defined constriction. Bursa obscurely
three-lobed, the posterior ray being slender and bifurcated at the extremity, its inner
branch extending to the border of the membrane. Spicules short, nearly equal. Tail
of the female long and pointed, the tip being subulate, and armed with a short, filiform,
transparent, and finely pointed appendage. Anus situated at the base of the tail.

Length of the male 8 lines.
Length of the female 10 lines.

Hob. Intestines of Eleplias indicus.

On the 24th of August 1870, I discovered large numbers of these small worms in the
intestinal canal of the Circus Elephant already referred to. To the naked eye they look
like ordinary thread-worms ; but a band-lens is sufficient to demonstrate their strongyloid
character. The parts about the mouth were not well made out ; but my representation
of the head is perhaps sufficiently characteristic (PL XXIV. fig. 4).

In its retracted condition the tail of the male worm, when its end is directed towards
the spectator, presents a distinctly hexagonal outline, the incurved points of the rays of
the hood creating six well-marked angular projections. The appearance in question is
at once both striking and delusive. "When the hood is flattened or artificially expanded,
then an oval or more or less heart-shaped outline results (PL XXIV. fig. 5). After
repeated examinations I constructed the following diagram. The anterior ray (a) is
double, and has the divisions closely applied ; it is bent forward Fi^. 5.

towards the axis of the body. The antero -lateral is simple, normal,
and also bent forward. The middle ray (c, d) is dichotomous,
and directed at a right angle to the axis of the hood. The postero-
lateral ray (d) is simple, long and narrow, but does not reach the
margin of the membrane. The posterior ray (e) is characteristic of
the species, being long, narrow, and divided near its end into two

mi 1 ! • t • n i.i Diagram of the hood and rays

unequal portions. The inner branch is nearly twice as long as the 0f Dochmius Sangeri
outer, and extends to the border of the membrane. When the (x 30 dliim')-

hood is expanded the end of the posterior ray is widely, because divergently, separated

DE. T. S. COBBOLD ON THE PAEASITES OE ELEPHANTS. 237

from its fellow. The third lobe of the bursa is deeply notched in the middle line, thus
imparting to the hood a spuriously four-lobed figure. The spicules are short, rather
broad at the base, and pointed at the tip or lower end. When the tail of the male
is viewed either from the front or from behind, the spicules display the not un-
common V-shaped outline.

The tail of the female worm presents well-marked characters (PL XXIV. fig. 0). In
addition to those mentioned in the diagnosis of the species, the uniform and rather
sudden narrowing of the body in this region, together with the presence of a con-
spicuous and somewhat gaping anal cleft, would render the future recognition of this
species a very easy matter, even if no male were procurable. In some specimens the
neck in either sex is much more contracted than in the particular example which I have
figured.

7. Filaria Smituii, sp. nov. (Plate XXIV. figs. 7-10.)

Body smooth, with very fine transverse stria?, filiform, and of equal thickness through-
out. Head furnished with two rows of translucent auriculate appendages, supporting a
few minute papilla?. Mouth simple, surrounded by four small papilla? and numerous
slight elevations exterior to the papilla?. (Esophagus of great length, without bulb or
constrictions. Tail of the male short, bluntly pointed, with winged appendages sup-
ported by eight stalked papilla?, four on either side. Spicules long, slender, unequal.
Tail of the female short, bluntly pointed, usually directed backwards. Anus sub-
terminal at the base of the tail.

Length of the male J of an inch.
Length of the female ^ of an inch.

Hab. Coats of the stomach of Elepltas indicus.

The delicate little worms thus named cannot be confounded with the larger species of
nematode already stated to occupy tumours in the walls of the stomach of the Elephant.
Apart from the relative disparity of the worms themselves, the parasitic growths or
abodes containing them are of much smaller size. The tumours enclosing the Filarice
barely exceed the size of an ordinary filbert nut, presenting also an oval flattened figure.
As in the larger formations, however, each tumour displays a single round aperture com-
municating with the cavity of the stomach. The interior structure of the growths does not
depart in any marked degree from the alveolar or multilocular character of the abodes of
Strongylus foliatus ; but a relatively far greater number of worms were found to occupy
the "abodes" of this species. I have named the species after the veterinary surgeon
through whose instrumentality I had an opportunity of making the " find." Well
enlarged, and seen in profile, the head shows sculpturings which together form a double
festooned border, whilst within the concavity of each ear-shaped fold of integument
there are minute papilla?. In the neighbourhood of the mouth there are also other
prominences, as well as special papillary elevations which probably serve as taste-
organs (PL XXIV. fig. 7).

In consequence of their larger size, the females are easily recognized. The actual
position of the reproductive outlet evaded my observation ; but the diagonal direction

238

DE. T. S. COBBOLD OX THE PARASITES OF ELEPHANTS.

Fie. 6.

Outline of the tail of a female

Filaria Smithii : a, anus

( X 30 diam.).

of an interior tube (which I took to represent the empty vagina) led me to infer that
the outlet is placed about yjjo of an inch below the point of junction of the oesophagus
with the chyle-intestine. The oesophagus itself, in the female,
measures ^ of an inch in length, and only ^ro m breadth. I
subjoin an outline of the female tail. It is here seen strongly
bent backwards. Another striking character is that shown by
the ova. These minute bodies, linear-oblong in shape, give
an average diameter of only ^wo °f an mcu from pole to pole.
Each female worm contains a prodigious number of eggs, pro-
bably not less than 100,000, exclusive of the unimpregnated
ovarian ova. The perfect ovum always exhibits an embryo
folded within its excessively delicate chorional envelope. This
embryo itself is thicker in front than behind ; hut it is bluntly

pointed at either extremity. Even at its thickest part the transverse diameter of the
body of the embryo scarcely exceeds the giuJo of an inch.

The head of the male forms a miniature representation of that of the female worm.
The difference in thickness in the two sexes is considerable, the body of the female
measuring xoo ail(i that of the male about yio 0I" an inch. Seen from the ventral
surface, the contour of the tail of the male presents a lanceo-
late outline and spatulate form. It displays eight papillae
furnished with long stalks (PL XXIV. figs. 8-10). On each
side three of the papillae are rooted or arise from above the
level of the anal opening, and one originates from below the
anus. I did not observe any sessile papillae beneath the
stalked papillae. The spicules are always visible, whether par-
tially (figs. 8, 9) or wholly retracted (fig. 10). They are
strikingly narrow, the anterior spicule being almost twice the
length of the posterior. The length of the shorter spicule
scarcely exceeds the rhj> whilst the anterior one
diameter of fully g^ of an inch.

Fie. 7.

gives a long

Outline of the tail of a male

Filaria Smithii; a, anus

(X30 diam.).

8. Ampeistoma Hawkesii, Cobb.

A. Hawkesii, Cobbold, The Veterinarian, Oct. 1875; ibid. Nov. 1875; also in Treatise on Parasites,

1879, p. 395.
A. Collinsii, var. Starileyi, Cobb., Parasites, 1879, p. 357.
Masuri or Mussoorie of the Hindoo Mahouts (Gilchrist, Hawkes, &c.).

Body of a pink colour, smooth, plano-convex, finely wrinkled transversely, bluntly
pointed and contracted in front, broadly rounded behind. Head surrounded by a few
regular but not well pronounced folds, armed with numerous short and extremely
minute warty papillae. Mouth terminal, circular. Ventral surface often slightly
depressed near the centre, forming slight prominences on either side. Caudal sticker
placed well forward, rather large, circular, with a broad lip and smooth concavity.

DE. T. S. COBBOLD ON THE PAEASITES OF ELEPHANTS.

239

Fig. S.

Outline of Aphistoma HawJcesii
(X4 diam.).

as the finder states, a

Reproductive papilla small, situated nearly midway between the mouth and upper
margin of the caudal sucker.

Length, on the average, f of an inch ; the longest specimen -~^ of an inch.
Breadth \ of an inch.

Hub. Large intestines of Elephas indicus.

In my general treatise, and likewise in the pages of the ' Veterinarian ' ante-
cedently, I have recorded the occurrence of this entozoon ; but
hitherto it has, zoologically speaking, remained undescrihed.
My first acquaintance with it dates as far back as June
1875. Towards the close of that month I received from Lieut. -
Colonel Hawkes a set of parasites for identification. Attached
to the Madras Staff Corps, Colonel Hawkes forwarded the
specimens from Secunderabad, at which station there had been
a recent and rather serious mortality amongst the Elephants.
The cause of that mortality will be subsequently considered
when I treat of the injurious effects of the parasites already
described. Having explained by a letter to the donor that
these parasites were new to science, I provisionally named
the species AmpMstoma Hawkesii in his honor.

In the fresh state the parasites have a bright rosy tint, or,
" delicate flesh-colour." Probably this colour, more or less pronounced, is invariable
amongst the Amphistomes ; at least such has been the case with all the forms that I have
examined in the fresh state. Similarly some Amphistomes received by me from an
Indian horse were reported by the finder as having a brick-red colour when removed,
post mortem. Here it is desirable that I should speak more fully of these equine
Amphistomes. They were found in the year 1875, by Veterinary-Surgeon Collins, of
H.M.'s Indian army, stationed at Simla. As equine parasites, however, they do not
appear to have been the earliest-seen specimens. None, indeed, had hitherto been
scientifically described as coming from the horse ; but my attention was subsequently
called to a set of specimens preserved in the private collection of Professor Simonds, at
the Royal Veterinary College. As since stated in my latest work, the parasites be-
longing to Simonds were at the time regarded by me as mere varieties of the species
found by Vet. -Surgeon Collins. I accordingly spoke of them as AmpMstoma Collinsii,
var. Stanleyi. I employed the variation-term in order to give expression to the circum-
stance that Professor Simonds received his specimens from Vet. -Surgeon Stanley, who
also obtained the Amphistomes in India. The date of Mr. Stanley's ' find ' has been
lost; but from Professor Simonds's recollection it would be about the year 1865.

I have recorded these few facts respecting the equine Amphistomes because, after a
recent and more careful study of the worms, I have come to the conclusion that the
parasites found by Mr. Stanley, though comparatively large, are really identical with
the particular Amphistomes I am now describing from the Elephant. At present,
however, I do not pronounce definitely respecting Collins's Amphistome, which will
probably be found to remain as a good species. In the horse dissected at Simla,

SECOND SERIES. — ZOOLOGY, VOL. II. 33

240 DE. T. S. COBBOLD ON THE PARASITES OF ELEPHANTS.

Mr. Collins reckoned that there were ahout a thousand of these parasites lodged in
the colon.

9. Amphistoma ornatttm, sp. nov.

Body of a pink colour, plano-convex, hut slightly concave anteriorly, pointed in front,
hroadly rounded off behind, and furnished with inconspicuous
transversely disposed rugae. Head and neck more or less pro- Ko. y

nounced. Mouth surrounded by circular folds, which are
armed with prominent, regularly arranged, and bluntly pointed
papillae. Ventral surface marked by lateral projections, which
fold regularly, sometimes showing a V-shaped groove at the
base of the neck. Caudal sucker subtcrminal. Reproductive
papilla large, and placed well forward.

Length \ to -A of an inch. . L..

° * Outline of Amphistoma ornatum

Breadth -^ of an inch. ( x 5 diam.).

Ilab. Intestines of Elephas indicus.

On the 24th of August, 1876, as already mentioned, I obtained a variety of parasites
from one of the herd of performing Elephants that perished at Sanger's Circus. Amongst
the parasites were quantities of amphistomes of various shapes and sizes. As all of them
were found in the same general situation, I at first concluded that they were of the same
species as those sent from India by Colonel Ilawkes. However, subsequent investiga-
tion led to the conclusion that, although we obtained two distinct forms of Amphistome
from the Circus Elephants, neither of them corresponded with the species received from
Southern India.

When I removed the parasites from the colon, they had the characteristic bright pink
colour, and were so transparent that a pocket lens was sufficient to show the general
arrangement of the internal organs. These will be referred to when dealing with the
next species.

I separate this form from Ilawkes's Amphistome chiefly on account of its smaller size,
its more conical figure, its large reproductive papilla, its slightly concave ventral surface,
and, more particularly, on account of the presence of papilla) in the neighbourhood of
the head. These papillary eminences are so conspicuous that with a magnifying power
of three or four diameters their projecting points are rendered visible.

10. Ampuistojia papillatum, sp. nov. (Plate XXIV. fig. 11.)

Body of a bright pink colour, smooth, conical, bluntly pointed in front, broadly
rounded off behind, with fine and regularly disposed transverse ruga) forming distinct
rings in the region of the head. Caudal sucker subterminal, very large, its cup being
armed with numerous large fungiform papillae, closely set, and regularly disposed over
the entire surface of the concavity. Reproductive papilla placed well forward.

Length £ to \ of an inch.
Breadth \ to \ of an inch.
Sab. Large intestines of Elephas indicus.

DE. T. S. COBBOLD ON THE PAEASITES OF ELEPHANTS. 241

This species is the smallest of the Amphistomes hitherto obtained from Elephants.
Regard being had to the occurrence of large papilhe within
the cup of the caudal sucker, no one could mistake it for lg'

either of the two previously described species. LTowever, it
is just possible that the somewhat similar Amphistome from
the Tapir, described many years back by Diesing, may turn
out to be either the same or a mere variety. Dr. C. M. Die-
sing's Amphistomct asperum differs in form, the anterior part
of the body being much constricted, whilst the caudal sucker u ^°H! ('^5 foam) /*"
is wider and terminal *.

The enlarged view that I have given of this worm was taken from a perfectly
fresh specimen (PI. XXIV. fig. 11). In examples preserved in spirit the aperture
of the sucker gapes more widely, and consequently looks larger. In the fresh state
one not only sees the digestive organs through the transparent envelope, but also the
testes, part of the uterus, the branching ducts of the vitellaria, and even also their
terminal sacs. The latter are so small that I have not attempted to represent them ;
moreover, except as regards position, neither these nor the other internal organs offer
any striking peculiarities. In this species the vitellary glands are limited to the sides of
the body ; but in Hawkes's Amphistome the sacs and their ducts seem to underlie the
whole surface of the body. From hardened specimens I obtained fine transverse sections
in which the disposition of the muscle-fibres and general stroma of the body were well
displayed.

As a believer in the common origin of all genera and species, of whatever zoological
type they may happen to be, I was not unprepared to find representative and parallel
forms of parasites in different animals. Thus it seems to me that the two Amphistomes
known to infest the Tapir represent two of the Amphistomes found in Elephants. On
the one hand my Amphistomct papillatum comes near to A. asperum, whilst on the other
my A. Rawkesii comes near to A. pyriforme. The Amphistomes of the American
Tapir were obtained by Natterer in Brazil in 1829, and were described by Diesing in the
Museum Annals already quoted. The Vienna helminthologist had not then lost his
eyesight ; and his elaborate memoir convincingly proves that modern anatomists, notwith-
standing our improved methods, have added little to our knowledge of the organization
of the trematode worms.

The fungiform papillae within the sucker of Amphistomct papillatum forcibly call to
mind the more striking processes lining the ventral disk of the amphistomoid worm
which I have called Gastrodiscus Sonsiuonis. In this remarkable genus, established as
such by Lenckart, the gastric papilla; display each a central depression at the summit ; and
I have therefore spoken of them as suckerlets. On this account, also, I pointed to their
close affinity with those seen in the equally singular genera Notocotylus and Aspidocotylus,
as established by Diesing. The possession of the small suckers in question, shared as it

* " Ncue Gattungen von Bmnemviirmen, ncbst cinem Nachtrage zur Monographic der Amphistomen." Annalen
des Wiener Museums, 1839, S. 236.

33*

242 DK. T. S. COBBOLD ON THE PARASITES OF ELEPHANTS.

is by all the three aberrant genera, seems to imply that their position amongst the
Amphistomes is somewhat doubtful ; and I am informed by Professor Leuckart that a
recent anatomical investigation of Gastrodiscus, made by one of his pupils, tends to
confirm this conclusion. Be that as it may, the simple papilla? within the caudal
sucking-disk of Amphistoma papillatum are evidently of the same character as those
found in A. asperwni. Diesing describes them as little prominences (Erhabenheiten).
As contrasted with the integumentary papilla? so frequent on the general surface of the
body of Amphistomes the papilla? placed within the disk of A.papillatum are remarkably
large and altogether distinctive.

11. Fasciola. Jacksoni, Cobb. (Plate XXIV. fig. 12.)

Fasciola Jacksoni, Cobbold, Quart. Jouru. Micr. Science, Jan. 1869; Supplement to Entozoa, 1869,

p. 80.
Distoma hepatkum, Jackson (non descriptum), Boston Museum Catalogue, 1847, p. 317.
Distomum elephantis, Diesing, Syst. Helm. Bd. ii. p. 560; Revision dcr Myzelmintlien, 1858, S. 354.

Body fiat, orbicular, often folded towards the ventral surface, smooth to the naked
eye, but armed throughout with numerous excessively minute dermal spines, which are
larger above than below. Oral sucker terminal, small. Ventral sucker large and well
forward. Reproductive papilla in the middle line, and placed considerably above the
upper lip of the acetabulum ; intromittent organ of great length. Digestive apparatus
branched, its ramifications ending in ca3cal terminations, which occupy nearly the whole
extent of the body internally ; oesophageal bulb distinct.

Length \ to f of an inch.
Breadth \ to \ of an inch.

Hab. Biliary ducts and duodenum of Elephas indicus.

As has been already stated, Dr. von Linstow has fallen into the error of assuming that
three species of fluke infest the Elephant. They are, in reality, all one and the same
species. Although long previously discovered, this worm was first recognized as a good
species by myself in the year 1869. Its discoverer, Dr. Jackson, gave no description of
the worm ; and it is clear, from what Dr. Fitz has said, as well as from the entry in the
Boston Catalogue, that Jackson supposed that he had encountered the common fluke of
ruminants in the Elephant. The description orginally given by me in the journal
above quoted differs in no essential particular from that now offered. I have merely
emphasized and extended certain characters.

As causing disease, and zoologically speaking, this entozoon is the most important of
the whole group of parasites infesting Elephants. Now that this fact is becoming known
to those interested in the preservation of the domestic pachyderms, more attention is
likely to be paid to the source and distribution of the parasite. In order, therefore, to
clear away certain misconceptions that have already arisen, it is desirable to restate the
facts of discovery.

As remarked in my recent work, the first specimens of Elephants' flukes received in
this country were transmitted by Vet.-Surgeon Thacker, of the Madras Army. They
were, as he expressly remarks, " forwarded for classification." The specimens, indeed,

DE. T. S. COBBOLD ON THE PAEASITES OF ELEPHANTS. 2VS

were brought from India by Dr. Hugh Cleghorn, who conveyed them to the late Dr.
Baird, who, again, honourably carried out the wishes of the donor by handing them
over to mvself. At the Norwich Meeting of the British Association in 1868 two of the
flukes were exhibited for the first time in this country ; and I then stated that they
" were identical with certain flukes long previously obtained from the duodenum and
liver-ducts of an Indian Elephant that died in America." I further stated that, although
the original ' find ' was carefully preserved in the Boston Museum, the worms had never
been properly described. In the summer of 1868 fifteen specimens of fluke, removed
from one or more Burmese Elephants, had been forwarded to Professor Huxley from
Rangoon, and the ' send ' was accompanied by a statement to the effect that the para-
sites were the cause of an extensive and fatal disease in Burmah. Through the kindness
of Professor Huxley, I was permitted to make use of his specimens for the purpose of
comparison ; and thus it became evident that his specimens and mine were of the same
species. I then published the original description already referred to ; and it was after-
wards repeated in the supplement to my elementary treatise on the Entozoa. The
second description of the worm was accompanied by an illustration, in which the general
plan of the mode of branching of the digestive canals was given on a reduced scale.

In his well-known systematic work, C. M. Diesing at first recorded the American
' find ' in his list of doubtful species ; but in his latest revision he formally recognized the
fluke as a good species. No description, however, was added. That revision was pub-
lished in 1858. Three years later, although I had not then seen any specimens, I
recorded Jackson's fluke as a Distoma, in the Synopsis of the Distomidte published in the
Linnean Society's Journal (1861). These references practically exhausted the literature
of the subject up to the year 1873, when I again had occasion to speak of these flukes.
(' Manual of the Parasites of our Domesticated Animals,' p. 13). Some of the specimens
received from Professor Uuxley were afterwards added to the entozoal series of prepara-
tions contained in the Uunterian Museum.

I think it necessary to explain the reasons for altering Diesing and Jackson's nomen-
clature as applied to this species, and all the more so since some helminthologists persist
in ignoring the generic term originally employed by Linnams in connexion with the
common fluke of Bmminants. "Whilst (excluding the Amphistomes and their allies) I
recognize the genus Distoma of Betzius as applicable to all the two-suckered flukes that
possess a simple unbranched dichotomous gastric organ, I think it unfortunate that my
proposal to retain the Linnean generic title, Fasciola, for such true flukes as possess a
dendriform or branched gastric organ has not been fully accepted. This structural
peculiarity is eminently characteristic of a few forms of fluke, and gives us a well-marked
type. Besides the species under consideration, only two other flukes are known to possess
the dendriform gastric apparatus. We have not yet even encountered any transition
types (in this particular), although, on evolutionary principles, one must conclude that,
if not now existing, flukes with their digestive tubes much less branched must have
existed. The only three fluke parasites yet shown to possess this remarkable dendriform
system of gastric canals are the Elephants' fluke {Fasciola Jacksoni), the common fluke
of Ruminants (F. hepatica), and the fluke found by me in the Giraffe (F. gigantea).

244 DE. T. S. COBBOLD ON TILE PAEASITES OF ELEPHANTS.

In 1864 I went fully into this question, and endeavoured to show that, at least as regards
the common fluke, the priority of the genus established hy Linnoeus was indisputable.
The genus Fasciola was accepted and recognized some twenty years hefore the genus
Distoma was proposed hy Retzius, and more than thirty years before it was adopted by
Zeder and his followers. (For dates, &c, see my ' Entozoa,' 1SG4, p. 149 ; also ' Para-
sites,' 1879, p. 15).

As furnishing interesting evidence of the value of cooperation in zoology, I may
mention that in one of my earlier notices I appealed to American naturalists, requesting
them to examine the original specimens of Fasciola Jaclesoni preserved in the Boston
Museum. This appeal met with a suitable response at the hands of Dr. R. H. Fitz, who
afterwards published the results of his anatomical inquiry. In regard to the original
find, we are told that "Dr. Jackson preserved a large number of flukes, and generously
placed them at the disposal of Dr. H. P. Quincy, avIio made and presented to the Warren
Museum a series of admirable preparations." These preparations consisted of longi-
tudinal, transverse, aud other sections, stained in carmine, and rendered transparent in
the oil of colours. Dr. Fitz expressly states that a study of them satisfied him that my
" surmise " as to the true character of the species they represented was correct ; and I
am glad to perceive that so able an observer did not hesitate to accept the generic
nomenclature that I had adopted.

The anatomical investigation made by Dr. Fitz resulted in not a few clear additions to
our knowledge, especially as regards the sexual organs. There is a very close corre-
spondence between the arrangements of the reproductive parts of this species and those of
the common fluke. In some particulars Dr.Fitz's observations and my own are at variance.
Thus Dr. Fitz appears to think that I have erred in attributing to Fasciola Jacksoni the
possession of integumentary spines. In my original diagnosis I affirmed that the body
is " armed with minute spines." I still adhere to that matter-of-fact statement ; but Dr.
Fitz says : — " Instead of finding the body armed throughout with minute spines, the
cuticle covering the abdominal surface presents a series of ridges pointing backwards,
and extending the entire length of the animal. Traces of a similar formation are present
on the dorsal surface or the neck ; elseiohere the back is smooth." The italics are mine.
I am sorry to correct Dr. Fitz. The little spines not only
exist, but when the skin of the parasite is examined by direct s"

light under a 2-inch objective, their equidistant and otherwise
regular arrangement becomes obvious. Under higher powers
the spines are measurable, giving an average of y^j to -jirs °f
an inch in length, their basal diameter being less than -^hro
of an inch. It must be admitted that their size is insignificant
when compared with the similar dermal spines of the common

n , TT,.77 , . , i . -i • i Eight cutaneous spines of

nuke. In Fasciola hepatica the spines are about twice as long D Fasciola Jacksoni

and thrice as broad. The differences therefore are sufficiently ( x 2^0 diam.).

striking to be emphasized in any diagnosis of the species ;

and, further, it may be said that they are perfectly independent of the ridges or folds of

the skin described by Dr. Fitz.

DE, T. S. COBBOLD ON THE PAEASITES OF ELEPHANTS. 215

On other grounds than those already stated, the peculiarities of the digestive channels
call for remark. Although drawn to a large scale (PL XXIV. fig. 12), I was not ahle to
represent the entire set of canals. The branches coming off from the main trunks of
this dendriform stomach, if such it may he called, are faithfully represented, both as
regards number, size, position, and extent of branching. In no two specimens, however,
is their form and distribution absolutely identical, though in all these flukes the general
arrangement of the canals is the same throughout. The specimen on which I chiefly
relied for an interpretation of the facts was injured in the mounting ; and in drawing that
specimen (without reproducing the injuries), I have derived much help from an outline
camera sketch made for me by Dr. John Foulerton. In spirit-specimens the natural
green tint or bile- stained appearance of the fresh worms is not entirely destroyed.

In the only hitherto published representation of the worm, I have purposely reduced
the digestive apparatus to the lowest possible expression consistent with the general
plan ('Entozoa,' Supplement, p. 19). That plan epitomizes, as it were, the larger
illustration now given ; yet neither the one nor the other fully expresses the degree of
complexity Avhich the secondary and tertiary ramifications of the channels exhibit. My
drawing represents the tubes and their csecal ends as seen from the back of the worm.
Though unaccompanied by any figure, Dr. Fitz's description closely corresponds with
what little I had already observed and written on this point ; but as he pushes the details
much further, and as many of his observations concur with what I have recently ascer-
tained to exist, I prefer to quote his accurate record of the facts, as follows : — " Just
beyond the terminal oral sucker is a voluminous bottle-shaped muscular pouch, the
pharynx, with which a short tube, the oesophagus, communicates. The latter has a
crenated outline, as if capable of a considerable elongation. This tube bifurcates at its
lower end ; the two short branches thus formed pass respectively to the sides of the
animal, and enter the main intestinal canals, which extend, one on either side, almost
the entire length of the body. These canals are widely separated from each other
until they appproach the tail ; and between them lie the sexual organs both male and
female, towards the front — the remaining space between, behind, and around being
filled in by the yolk-glands or tubes and the caical pouches of the intestine. The
latter arise as primary, secondary, and even tertiary divisions, from branches given off
more or less alternately from the main canals ; and their general direction is backward,
Those running towards the median line are short, with comparatively few branches,
while the external tubes are considerably longer, often bifurcating before the secondary
and tertiary canals arise. The largest number of caecal pouches are thus observed in the
outer portions of the parasite. Another peculiarity of the external canals is, that their
secondary branches are almost invariably given off from the posterior aspect of the tubes
from which they arise."

In describing the other internal organs, Dr. Fitz supplies some interesting particulars ;
but it is not always clear whether he is dealing with the facts as he actually found them
or with appearances which he correctly inferred to exist. Although in my own speci-
mens I obtained clear evidences as to the general distribution of the reproductive organs,
and especially of the yolk-glands, yet I could neither trace out the ducts of the vitellaria

2^6 DE. T. S. COBBOLD ON TIIE PAEASITES OF ELEPHANTS.

nor even the vaginal prolongation of the uterine coils. Their position, however, can he
perfectly well understood from the position of the parts actually seen. I have not
attempted to figure the yolk-sacs ; hut the illustration shows a portion of the uterine coils
and their contained ova. As in the common fluke, the eggs are oval, and furnished
with a lid at one end of the shell. They are comparatively small, affording an average
of _i_ 0f ari inch in length and -g^y in breadth. Dr. Fitz doubts the existence of any
true reproductive papilla, but speaks of a pocket or depression midway between the
pharynx and the ventral sucker. Certainly the depression referred to is occasionally
very marked ; but in the centre of this depression, even in the retracted state of the
cirrhus, there is a papillary eminence, which may be seen with an ordinary hand-lens.
The exserted cirrhus was not seen by Dr. Eitz ; consequently he could not confirm my
statement as to its remarkable length. As shown in my figures, this whip-like organ
nearly equals in length the transverse semidiameter of the body.

In my previous account of this worm the invariable distention of the digestive tubes
by inspissated bile derived from the host has been remarked. Dr. Fitz noticed the same
thing in Jackson's specimens. Neither in the common fluke nor in the fluke of the
Giraffe has this natural injection been found to be so perfect. For anatomical purposes
it renders any artificial distention of the ducts quite superfluous.

Note. — The foregoing observations complete my account of the helminths properly so
called. Although, without question, as regards their zoological significance and the
production of injury to Elephants this class of parasites holds the chief place, yet I
cannot omit all notice of the Arthropodous parasites. Doubtless much more remains to
be done in reference to the insects and mites that attack Elephants either as parasites
or as tormentors. The burden of this research, however, must fall rather on the ento-
mologist than on the helminthologist. Arguing from what is already known respecting
the ectozoal Arthropoda of other pachyderms, I should suppose that we have here a wide
field for further discovery. When it is considered how large a number of cestridean
larva; take up their abode in herbivorous quadrupeds, to say nothing of the ectozoal
Anoplura and Trachearia, it seems strange that not more than three or four arthropodous
parasites are at present known to infest Elephants. These I proceed to notice.

12. Gastrophiltjs elephantis, Cobb.

(Estrus of the Elephant, Cobbold, Catalogue of Entozoa in the Museum of the Royal College of

Surgeons, 1806, p. 24.
(Estrus elephantis, Steel, Med. Examiner, 1878, p. 886.

Head and neck of the larva turned forwards, prominent, spinous ; epicranial lobes
armed at the tip. Maxilla? large ; mandibles concealed. Somites of the body all more
or less armed, the hook-circles of the middle somites supporting five rows of spines,
of which the two upper rows are more conspicuous in consequence of the relatively
greater size of the separate spines ; penultimate somite with very few spines, and the
last somite nearly destitute of them. Caudal lobes furnished with six prominent

DE. T. S. COBBOLD ON THE PAEASITES OF ELEPHANTS.

247

papillae, two in front and four behind, numerous spines occurring in the space between
the papilla} and the border of the oval caudal plate.

Length of the larva f to 1 inch.
Breadth -^6 of an inch.

Sab. Stomach of Loxodonta africana and ~Elcphas indicus.

When in the year 1865 I rearranged the Entozoa preserved in
the Hunterian Museum, I added a few new forms of (Estridean
larvae which had been presented to me by the late Mr. Andrew
Murray and by Dr. (now Sir John) Kirk. Amongst the ' bots ' obtained
during Livingstone's Zambesi expedition were several taken from the
stomach of a female Elephant shot by Dr. Kirk himself. Some of
these, with others from a Hartebeest, remain in my possession. Those
that I placed in the Hunterian Museum were entered in the Cata-
logue as donations from the eminent naturalists to whom I stood
indebted.

Although hitherto undescribed, probably many Indian as well as
African sportsmen have encountered these parasites in slaughtered
Elephants. Be that as it may, no one appears to have had any
opportunity of comparing and identifying the * bots ' derived from
the widely dissimilar proboscidean pachyderms of the two countries
in question. In the autumn of 1878, one of the victims of the

,',.,. Bot of Gastrophilus

epidemic at Sanger's Circus fortunately supplied me with the means elephantis ( x 3 diam.).
of comparison. The longest of Dr. Kirk's specimens barely measured -x% of an inch,
whereas two of the specimens removed from Sanger's Elephant exceeded an inch in
length. At first sight probably few would hesitate to separate the bots of the Indian
and African Elephants, thus regarding them as the larval representatives of different
species of gad-fly. However, a careful examination of the somites, of the arrange-
ments and proportions of the spines, of the papillae, and of other characters, has con-
vinced me of their identity. Not only so ; the African and Indian Elephant bots,
though identical in themselves, differ from the bots hitherto found in the stomachs of
other animals.

The Indian Elephant, a male, which yielded the specimens obtained in this country,
was examined by Veterinary-Surgeon Erederick Smith, R..H.A., and by Mr. Steel, to
whose memoir on the circus-epidemic allusion has already been made. Speaking of the
post-mortem results, Mr. Steel says: — "The very numerous bots (CEstrus elephantis)
were found in the stomach adhering to the lining membrane. They Avere, as a rule,
larger than the bots of the horse, and exhibited a greater degree of ferocity. As soon as
removed from the membrane they struggled violently ; and several grasped my finger
firmly with their booklets. The female animal died shortly afterwards ; but we had not
an opportunity of making a post mortem examination of her. A fourth Elephant died
about a week after delivery" [to Mr. Sanger]. Mr. Steel then goes on to speak of the
opinion which he gave in a court of law as to the cause of these animals' deaths ; but

SECOND SERIES.

31

248 DE. T. S. COBBOLD ON THE PAEASITES OF ELEPHANTS.

it is evident that the ' bots ' were little, if at all, concerned in bringing about the
fatal issue.

These bots are perfectly distinct from those obtained from the stomachs of other large
animals. Not many years back Mr. Spooner Hart, V.S., of Calcutta, sent me some fine
larva? taken from an Indian Rhinoceros. They are evidently identical with the
Gastrophilus rhinocerolis obtained by Prof. Owen from the stomachs of African Rhino-
ceroses (B. bicomis and R. simus), and cannot be confounded with the bots of the
Elephant. As regards size, the bots of Rhinoceros indicus quite equal the measurements
recorded by Brauer as characterizing the bots of the African Rhinoceroses.

Parasitic Ova. — In this place it is fitting to notice the curious kind of parasitism
believed either to result from or to have some immediate connexion with the deposition
of eggs of an insect.

At a meeting of the Zoological Society held on the 21st of February 1871, Mr. Sclater

showed some tusks of a female Indian Elephant, their surfaces being " corroded or

eaten away " near the gums. Egg-like bodies, probably those of a dipterous insect,

were found attached below the affected spot. According to Mr. Roden, who shot the

Elephant, the eggs were on the tusks at the time he killed the animal ; but " there

were no maggots in the grooves at the end of the tusks." To whatever circumstances the

injuries were due, it is clear that the corrosions were not accidental; for, as Mr. Sclater

pointed out, a well-known Indian sportsman had already called attention to the matter

in the columns of 'The Field.' Under a pseudonym (" Smoothbore") the sportsman

solicits information from zoologists, and remarks in particular on the circumstance that

the corrosions or injuries " have never been observed in the tusks of the male." In

confirmation, Professor Flower also called attention to the fact that the Hunterian

Museum contains a pair of tusks similarly marked. Thus, to whatever parasite the

appearances in question are due, it is clear that their occurrence is sufficiently common,

and that there ought to be no difficulty in procuring materials for further investigation.

Dr. Mas Schmidt, the Director of the Zoological Garden at Frankfort, has also referred

to the ova of insects on the tusks of Elephants. His well-known memoir is exhaustive

and elaborate as regards the general diseases of pachyderms ; but little is said about

their parasites. Not one of the entozoa of Elephants described by Baird and myself

appears to have come under Dr. Schmidt's notice.

Through the help offered by Professor Flower, I have had an opportunity of examining
the tusks preserved at the Royal College of Surgeons. In the Catalogue it is stated
that one was obtained from a young Ceylonese female Elephant. This tusk was pre-
sented to the College by Mr. G. H. K. Thwaites. The other tusk, presented by Mr.
Sclater, was removed from a wild and full-grown female Elephant killed at Malabar.
Both teeth display deep erosions at the surface, near the edge of the gum ; and in the
smaller specimen from the young Elephant the erosion is about a quarter of an inch
in depth. The eggs are deposited on the tusks side by side in single file, but not in
perfectly straight lines. They form gracefully curved groups, each group presenting a

DK, T. S. COBBOLD ON THE PAEASITES OF ELEPHANTS. 249

festooned appearance. On the small tusk, from which a large number of the eggs
had evidently been rubbed off, I counted 150 ova — that is, reckoning the perfect and
imperfect shells, as well as the impressions left by such as had been detached. The
counting of the eggs and markings on the larger tusk was a more tedious process ; but I
satisfied myself that at least 2000 ova had been deposited. The Assistant Conservator,
Dr. Garson, obligingly removed one or two egg-shells. I found that these ova gave an
average length of ^ of an inch. Separately they presented a linear-oblong figure,
being rather thicker at one pole than at the other. In a perfect
specimen the narrow end measured the -g1^ of an inch, the broad end being
about the ^ of an inch in diameter. From certain appearauces I am
inclined to think that the intra-ovular maggots make their escape by the
opening of the shell in a lid-like manner. Large numbers of the eggs,
have their shells open as if three fifths of the exposed surface had fallen
away after the manner of a lid. The external configuration shows a
rather sudden transition from the thicker to the thinner part of the egg,
which circumstance also favours this view. Considering, moreover, the
known facts relating to the form and contents of the eggs of gad-flies

Eij£j of <i Dii>tcron

generally, it appears to me not improbable that the ova in question have (x 20 diam.).
been deposited by the mature Gastropkilus elepkantis. It is a question which the
entomologist is alone competent to settle ; but I may observe, as favouring this view,
that since the tongue of the Elephant could not reach the ova if the gad-flies deposited
their eggs on the legs and shoulders (as obtains in the Horse), it becomes a necessity for
the gadfly of the Elephant to deposit her eggs either on the coarse hairs in the neigh-
bourhood of the mouth, or in situations similar to those actually occupied by these
dipterous ova. It is scarcely likely that the trunk of the Elephant would or could be
employed to detach ova measuring only the ^t of an inch in length.

Ectozoa. — There are two other arthropodous parasites known to infest the Elephant :
the one is a huge louse, the other a species of mite. Hcematomyzus elepkantis differs
from ordinary lice in many particulars, but it has been remarked by Piaget that the
reproductive organs resemble those of the genus Hiematopinns. Mr. C. H. Ptichter
described, in the pages of ' Science Gossip ' for 1S71, what, at the time, was sup-
posed to be " a new form of parasite," which he called Idolocoris elepkantis. This
insect, one line in length, was found upon an Elephant in Ceylon. According to
Walker, who defined the genus, it not merely has supplied us with the type of a new
genus, but with that of an entirely new family of the Hemiptera Heteroptera, coming
very near to the bed-bugs (Acanthidae). In the same publication (p. 231) Dr. Buchanan
White suggested the name Phantasmocoris for the genus ; but a subsequent writer (ibid.
p. 278) cleared up the difficulty by referring the parasite to Piaget's Hcematomyzus.

The excellent figures given by Richter and Piaget leave little to be desired; but
notwithstanding Piaget's explanation, the specific term longirostris would perhaps have

31*

250 DR. T. S. COBBOLD ON THE PARASITES OF ELEPHANTS.

been preferable. This Ectozoon is no longer a rarity, specimens having been exhibited at
a recent meeting of the Royal Microscopical Society and elsewhere.

I have nothing new to add respecting the mite of the Elephant. Its position is still
doubtful. By Eiirstenberg this Acarus was called Homojms elephantis, and by Gerlach
Symbiotes elephantis. According to Megnin, it is a nymphe adventive, or hypope, of a
variety of Tyroglyphus siro, which is very abundant in old forage. Now, if the views of
Megnin be correct, Homopus elephantis, considered as a good species, falls to the ground.
Not only so ; it cannot fairly be called a parasite proper to the Elephant, since its
occurrence on the only animal on which it had been found was purely accidental. The
original specimens which supplied the means of description were sent by Gurlt to
Fiirstenberg. They were found in immense numbers on the skin of an Elephant
recently packed with straw (empaille). It was referred by Fiirstenberg to the parasitic
Acaridse, but not to the true itch-mites, properly so called. As implied by his nomen-
clature, Gerlach had taken the latter view. After studying Megnin's criticism of the
whole Acarine question, one feels bound to acquiesce in his authority. On the question,
however, which he has recently raised respecting the dimorphism of the Cestodes, I
think M. Megnin's views are altogether at variance with any correct interpretation of
the facts known to result from experimental research.

So far as I am aware, no other parasites of Elephants have been discovered; but
without doubt several more species will in time be brought under observation, especially
from amongst the Arthropoda.

Practical Considerations.

Although much has been written on the diseases of Elephants, little has been said of the injuries
produced by parasites. Having, by correspondence, supplied Colonel II. P. llawkes and other residents
in India with particulars respecting the various parasites transmitted for identification, these gentlemen
have, in their turn, furnished me with the practical results of their observation and inquiry. What
I have already published on this head has been either extended or confirmed by subsequent correspond-
ence ; but the most considerable additions to our knowledge have been made by the valuable communi-
cation of Mr. John H. Steel, in which full details of the Circus cpizooty are described. If properly looked
after, Elephants in captivity ought to suffer less from parasites than their companions in the wild state.
We have not, indeed, any reliable facts on which to form an opinion as to the extent to which wild
animals are thus victimized. That, like their captive congeners, wild Elephants perish from ' fluke-rot/
there cannot, I think, be the slightest doubt ; for in the natural state they have the same or even
increased facilities for swallowing the parasitic larva;. Guided by the recent determination of Leuckart
respecting the intermediate or Molluscan bearer of the common fluke, there is every likelihood that the
mollusk harbouring the Cercarian larva of Fasciola Jacksoni is small, and possessed of amphibious habits.
Not impossibly more thau one species of mollusk is concerned in the intermediate office, although
I strongly suspect that in the case of Fasciola hepatica the representative Cercarian larv« limit their
temporary residence to the body of a single mollusk. If this be so, Limn<ea truncatula is the responsible
' host/ and the source of sheep-rot. Be that as it may, we shall never settle this question in connexion
with the flukes of the Elephant unless our Indian zoologists take up the study of parasites in the same
earnest spirit in which so many of them have advanced other branches of natural history.

Captain Williamson's ' Oriental Field Sports ' is one of the earliest works in which any thing is said

DE. T. S. COBBOLD ON THE PARASITES OE ELEPHANTS. 251

respecting the parasitic diseases of Elephants ; and what he states about the habits of these pachyderms
affords a clue to the source of fluke-infection. Of course, he wrote at a time when it was not known
that freshwater mollusks, especially those frequenting marshy grounds, were the responsible media of
fluke-infection. Captain Williamson says : — " The Elephant, like the Buffalo, delights in wallowing,
and never thrives so well as when he is allowed to visit a rapid stream, there to exercise himself in
swimming, as well as to lie immersed and cool himself." Further on he says : — " Elephants are much
troubled with worms, for the cure of which complaint the Elephant cats earth ; this acts both as an
absorbent and operates as a powerful aperient. Those who may have kept Elephants for years, and have
not attended to the minutiae, have yet to learn that the Elephant, being sensible of his malady, resorts
to this simple remedy voluntarily ; all the mohout has to do on such an occasion is to keep him from
eating other food, but to allow him abundance of drink. If the fajces be inspected there will be seen
an amazing number of moving objects, which so much resemble pieces of sugar-cane or of green bough
as to require some attention to distinguish. They are ordinarily about the size of a grain of oats, and
from their peculiar form no doubt prove particularly distressing and injurious to the stomach and
intestines." In making this brief extract I have omitted one or two words and synonymized another.
It is clear that Williamson was speaking of the Nematodes, and was probably unacquainted with the
fluke-disorder. Referring to the intestinal worms, he remarks the rapid decomposition of Elephants
after death, as a ground of apology for not having by dissection ascertained more precisely the true " seat
of this disease." He shows that the administration of the " solution of the Kallah-nimok, or bit-nobeu
(the quantity, perhaps, of four ounces to as many gallons or more of water)," is amongst the best
remedies that could be selected. Williamson says this di-ug " is well known as the most common
purgative in use throughout India." On the authority of Dr. E. J. Waring, I may state that it is
merely " an impure chloride of sodium, containing a varying proportion of sulphuret of iron." It is,
Waring adds, " a very popular Hindu nostrum."

On the subject of earth-swallowing, the statements of Williamson are quite in harmony with the
opinions held in India at the present day. As Mr. George Fleming has remarked, the same opinions
are recorded by Captain Forsyth. This sportsman says : — " Elephants arc very liable to intestinal
worms. They generally cure themselves when they \_i.c. the worms] get very troublesome, by swallowing
from ten to twenty pounds of earth. They always select a red-coloured earth for the purpose. In about
twelve hours after, purging commences and all the worms come away. When this occurs the hard
food should be stopped for a week, fodder only being allowed, and a ball of spices should be given every
day. Some Elephants will not eat earth when they require it, and they are considered a very bad lot in
consequence." Probably most of us have witnessed this earth-eating habit of which Indian writers
make so much. With others I have watched the same habit on the part of the African Elephants at the
Zoological Gardens in the summer time. It is one thing to observe a common-place fact, and quite
another to accept popular inferences. The credit for intelligent self-doctoring which the mahouts ascribe
to Elephants, is on a par with certain other tales that are told us respecting the wonderful self-culture
and instinct of these animals. Assistant Commissary-General Hawkes and other officers, both civil and
military, with whom I have conversed, seemed all more or less committed to this view. Speaking of the
Amphistomes as Masuri, Hawkes says : — " These parasites appear to be very generally present in the
Elephant. When their numbers are few, the host is probably not much inconvenienced ; but when
present in any great quantity they undoubtedly cause much irritation. When this is felt the animal
instinctively resorts to a simple and effectual remedy : he eats a quantity of earth, which purges him
thoroughly, and expels the Amphistoma. The Mahawats are of opinion that whilst the Elephant is
eating earth to relieve himself of the pests, the daily allowance of rice should be scrupulously withheld ;
and they say that if the rice, which is giveu uncooked, is eaten by the animal under these circumstances,
excessive purgation is induced which frequently results in death." Colonel Hawkes adds that the

252 DR. T. S. COBBOLD ON THE PARASITES OF ELEPHANTS.

Mahawat's name for the disease means ( fasting/ which expression " bears testimony to the generally
received notion of the necessity of withholding the rice when the animal is eating earth."

In 1841 Mr. Gilchrist published his often quoted ' Practical Memoir on the History and Treatment
of the Diseases of Elephants ; ' and in this work, without any attempt at scientific definition, he alludes
to the parasites termed Mussoorie. He also speaks of the Nematodes as Shotee, and likewise of the bots
or, rather, the disorder they create, under the term f Lemgum.' The rendering of native names in
English varies amongst Hindustani scholars ; and thus Hawkcs spoke of the Amphistomcs and Nematodes
as Masuri and Soorti respectively.

From the evidences I have collected it is clear that epizootics affecting Elephants are more or less
frequent in India and the East. The flukes received by Professor Huxley from Rangoon were obtained
from an animal which was only one out of many victims that perished from ' rot ' in 1867. A resident
in Burundi, under the signature of R. B., described this outbreak in a public journal. The Secunderabad
epidemic, of which I received full particulars from Commissary-General Colonel Ilawkes, was also para-
sitic, and, in my opinion, partly due to the presence of flukes and partly to the Amphistomcs. Speaking of
the post-mortem examination, where several kinds of parasites existed, Vet.-Surgeon \Y. S. .Adams says : —
" The large number of flukes in the liver and the intestinal parasites account in a great measure for
some of the symptoms shown; and these symptoms accord in many respects with those shown in
Elephants that died in Burmah during the epizootic in 1867." Mr. Adams adds that the liver-parasite
is the same as that which I had described as Fast-lulu Jacksoni; and in regard to the symptoms in the
Elephauts, he noticed especially the " refusal of food, standing with the mouth open, restlessness, and
puffiness about the head and shoulders." Speaking of another of the Secunderabad victims, the same
observer remarks : — " There were flukes in the liver, but in no great quantity; and the structure of the
liver was sound. Although not assisted by this case in attributing the mortality to parasitic origin,
I am strengthened in my opinion that the death of the previous Elephant was due to disease caused by
the presence of the fluke." Mr. Adams does not himself speak distinctly as to the presence or absence
of Amphistomcs in the Secunderabad outbreak ; but, as we have seen, Colonel Hawkes remarked both
upon their prevalence and the power of mischief that they had exerted in this epidemic.

In various published papers I have shown that Amphistomcs frequently infest horses in India ; and
in one case (examined after death, and reported to me by Veterinary-Surgeon F. F. Collins) it was
calculated that about a thousand of these parasites were present. " Nearly the whole of them were
situated close to the crecum, and were loose in the gut." Mr. Collins sent me thirty-three of the
specimens.

In the Secunderabad epizooty fourteen elephants perished. Writing to me in May 1875, Colonel
Hawkes said : — " Out of twenty-eight Elephants under my charge, no less than twelve have died within
the last sixteen months, whereas the average annual mortality has been hitherto only two per annum out
of thirty-eight on our establishment." My military informant, much to his credit, caused a post mortem
to be made in almost every instance ; and he supplied me with the results obtained and recorded by the
various veterinary surgeons. Two of these records (made by Mr. Adams) I have already quoted. Taking the
report as a whole, I am sceptical as to the accuracy of the deductions it embodies. It is clear that the
mortality was unusual, not to say unprecedented, and that therefore some special cause must have arisen
as a chief, if not as an exclusive, factor in its production. In my opinion, and even supposing the para-
sites were not the real cause of the mortality, the report is eminently unsatisfactory. Parasites or not,
here was a local and fatal epidemic ; and yet, as a result of the examination of thirteen carcasses, the
deaths are attributed to no less than eight different causes — inflammation of the intestines, inflammation
of the lungs and liver, splenic apoplexy, sunstroke, and so forth. Knowing, from long and special pro-
fessional experience, the scepticism which prevails in the medical world in respect of the injurious effects
of parasites, I cannot affect surprise at the conclusions formed by the veterinary surgeons in India. To
the mass of practitioners helmiuthological science and its practical issues remain a terra incognita.

DR. T. S. COBBOLD ON THE PARASITES OF ELEPHANTS. 253

Colouel Hawkes himself was evidently perplexed by the record of results. He wrote, saying : — " In
every case at which I was present flukes were found in greater or less numbers in the gall-ducts of the
liver ; and the Amphistoma was also as constantly present in the intestines, the Soorti (Ascaris lon-
choptera), contrary to the general experience of the elephant-attendants, being less frequently met with,
though, from its colour and its shape, it is not so easily detected among the huge mass of fieces as the
larger Amphistoma." There must, of course, be some misapprehension here, unless the Strongyles are
included in the general term of Soorti ; for the particular nematode above mentioned was not present in
in any of the bottles of specimens which I received from Secunderabad. " In some cases flukes were in
large numbers, apparently collected in bunches in the gall-ducts ; but disease of the liver did not neces-
sarily appear to follow, nor, where disease was actually present, did it seem to bear a constant ratio to
the number of flukes in the gall-ducts." Colouel Hawkes subsequently remarked upon the impossibility
of concluding to what extent the flukes were responsible for the outbreak, and the difficulty of singling
out any symptom as characteristic of their presence. This does not surprise me, considering the number
and variety of parasites that were admitted to have been present, not alone in the liver, but also in the
intestines. In one instance there was the characteristic swelling beneath the jaw, similar to that found
in sheep in the advanced state of ' rot,' when the animals are said to be suffering from " bottle-jaw."
This "suggestive" feature, as Hawkes himself called it, seemed in his view somewhat neutralized by the
fact that in " only the one other case (in which the parotid glands were swollen) the liver was more healthy
and had fewer flukes than any in the whole series " of dead Elephants.

In what has been called the English outbreak, affecting Sanger's performing Elephants, the injuries
appear to have been principally due the Amphistomes. In the animal examined by me, these Entozoa
were present in prodigious numbers. In all, five animals perished, one during the sea-voyage, and the
remaining four some time after their arrival. I feel assured that all the parasites found in these animals
were brought by their hosts from India. Mr. Steel records the symptoms separately observed in three
of the victims. The record is too long for quotation ; and, as one might expect, there is not much that
is distinctive about the symptoms. Speaking of the post mortem of a young female Elephant, Mr. Steel
states that the large intestine " was studded with very numerous, small, vesicular, blood-coloured para-
sites, about the size of grains of wheat. In addition, we found specimens of nematode worms inter-
mingled with the bowel-contents, some being found in the dilated portion of the bile-duct, where it
terminates at the duodenum. We found a few specimens of a thread-worm on the surface of an
incised part." By an anonymous writer in the ' Lancet ' it was stated at the time that this female
Elephant had suffered from " double plenritis ; " but, as Mr. Steel himself subsequently pointed
out, "the absence of a pleural sac has been proved to be a normal condition in this animal." It is
worthy of remark that in the case of the remaining patients the treatment adopted sufficed to expel some
of the parasites ; but from the first it seemed pretty clear that the animals were past all remedial help.
One of these young Elephants, a male, died suddenly. " In this case the parasites, both Amphistomata
and Ascarides, were much more numerous, the former existing almost in thousands." It was this
animal also which supplied the stomach-bots described in this memoir.

Considering the heavy losses thus incurred, it was not surprising that the owner made the matter a
subject of litigation on the ground that there had been a breach of warranty. Without giving Mr.
Steel's views in full, I may state that he shared with me the opinion that these animals were infested
by parasites at the time of their landing in this country ; and further, he believed that if it could not
be shown that they were the sole cause of the epidemic, the parasites at least very materially contributed
to bring about the fatal issue.

One passage in Mr. Steel's communication is perhaps a little too significant. He says : — " The
parasites in the bowels gave rise to irritation ; and the Amphistomes, like so many small leeches, were
withdrawing large quantities of blood from the animal ; the bowel-lesions were most marked where
these parasites were found." Mr. Steel seems inclined to set little store by the action of the Nematodes ;

254 DR. T. S. COBBOLD ON THE PAEASITES OF ELEPHANTS.

but as these turn out to be not Ascarides, but Strongyles, armed with more or less formidable teeth,
I am inclined to think that he has underrated their powers of mischief. On the other hand, whilst I
quite think with him that the Amphistomes, by their sucking action and consequent irritation, are
capable of giving rise to severe inflammatory lesions, yet I doubt their capacity for drawing blood in
the way Mr. Steel implies. If the Amphistomes really gorge themselves in this sucking process, they
must needs be continually detaching themselves for the purpose of ejecting the contents of the stomach.
In true leeches, as also in those Strongyles which have bloodsucking propensities, the presence of
an anal outlet enables them, whilst still attached, to discharge freely those constituents of the blood
that are not necessary for their nourishment.

In reference to the Ectozoa, I possess no sufficient data whereby to estimate their power of pro-
ducing mischief amongst Elephants. Judging from what happens in the case of other domestic animals,
it is certain that strict attention to sanitary matters will diminish the chances of invasion either from
external or from internal parasites. Mr. Steel speaks with evident satisfaction of the cleanliness and
care taken by the grooms at Sanger's Circus-stables, where the flooring was "subjected to frequent
disinfection." At the Zoological Gardens I have frequently made inquiries of the Head Keeper
(Scott) whose diligent search for Entozoa has hitherto been attended with negative results. This
speaks well for the cleanliness observed there ; and at the same time the absence of parasites shows
that the habit of earth-eating, which the African Elephants at the Gardens are wont to indulge in, is
not consequent upon the presence of Entozoa. This habit is not confined to any one class of animals.
My friend Mr. Arthur Folkard, during his residence in Ceylon, noticed that the Elephants of that island
were large earth-eaters ; and the late Dr. J. P. Rowe assured me that Australian horses and sheep
largely indulge in the same propensity. Dr. Rowe thought that the sheep thus sought instinctively to
relieve themselves from the irritation produced by parasites in the stomach. My own belief is that when
colic-symptoms are associated with a depraved appetite, the animals affected swallow mud, soft earth, or
sand indiscriminately. A remarkable instance was brought under my notice by Colonel Hawkcs, where
the accumulation of " sand or gravel " in the intestines of a military horse amounted to 14| lb. It was
supposed at the time to have resulted from careless feeding with impure grain or rather 'pulse;' but this
is impossible, as only two or three ounces of sand could be swallowed as the result of the daily meal. At
this rate, as Hawkes observes, "it would take from 77 to 116 days to accumulate 14| lb. In con-
sequence of this earth-eating habit, he adds (from memory) it was found necessary, some twenty years
ago " to remove the mud-walls round the picket enclosing some of the horses of a mounted corps. If
the conclusion formed by Commissary Hawkes and others be correct, all foul-feeding horses and elephants
must either be suffering from Amphistomes or some other kind of internal parasites. Be this as it may,
any sanitary measures which prevent the ingress of parasitic larva; cannot fail to prove beneficial. In
like manner any agencies which prevent a parasite completing its life-cycle must be equally effective.
Thus clean water and fodder given to Elephants in captivity prevents the possibility of their contracting
either masuri or suorti. I think it extremely improbable that any Elephant should contract internal
parasites in this country. Mr. Steel thinks otherwise, and that, as regards the Trematodes, they, or
rather their larva;, " might make shift with an English mollusk." It is not, however, so much a question
of making shift on the part of a Cercarian as it is a question of an allotted or appropriate territory. Some
Trematode larva; are limited to one host, and others to a few hosts. In all cases there is a measure of
restriction ; aud I think we must therefore look to one or more Indian mollusks as the legitimate and
aud only bearers of the Cercarian larvse of Fasciola Jacksoni. But, as before hinted, the final settle-
ment of this question is not yet awhile. Eveu with a large accession of labourers in this field of re-
search, the process will be tedious, when it is considered how long a time, aided by frequent observation
and experiment, helminthologists have taken to acquire the meagre results already arrived at in con-
nexion with the parasites of cattle ; it is evident that as regards the origin of, and transformations under-
gone by the internal parasites of Elephants, we are only on the threshold of inquiry. It may be said,

DE. T. S. COBBOLD ON THE PAEASITES OF ELEPHANTS. 255

indeed, that our efforts have not yet properly commenced. If this be the case, any thing like an exhaustive
knowledge of the development of the Elephants' helminthiases is a matter likely to be realized only in the
far future. There is not much done to encourage research in this special direction. Those who have
charge of Elephants have usually so little knowledge of zoology that their observations are either
common-place or altogether unworthy of credence. Even thoroughgoing naturalists shirk the subject.
Certainly the 'Challenger'' and other expeditions expressly fitted out to collect all kinds of natural-
history objects might have have accomplished much more in this direction had they been minded to
follow the example of previous explorers. The most interesting private contribution of parasites in my
possession I owe to the liberality of Mr. Darwin, he having obtained some Entozoa about forty years
ago during his travels in Patagonia. But I must conclude. If the publication of this imperfect memoir
should do no more than gain for the helminths greater consideration at the hands of naturalists, I shall
feel that my efforts in a really useful though unpopular cause have not been altogether fruitless.

APPENDIX.

Since the reading of this memoir took place, Mr. R. McLachlan, P.Pt.S., has referred
me to a paper written hy Dr. F. Brauer in the lGth vol. of the Transactions of the
Zoologico-Botanical Society of Vienna. It is entitled " Pharyngobolus qfricanus, ein
Oestride aus dem Rachen des afrikanischen Elephanten." Mr. McLachlan also sub-
sequently received a letter from Dr. Brauer, in which the latter savant refers to the
abstract of my paper already given in the Linnean Society's Proceedings. Not
unnaturally Dr. Brauer appears surprised that the parasite described by him has been
omitted from my list ; and he thinks it probable that his Pharyngobolus is either similar
to or identical with my GastrophUus elephantis. Dr. Brauer's inference is not correct.
The two bots found by BZerr E. J. C. Marno in the oesophagus of the African elephant
that died at Vienna are quite distinct. Their characters, as given by Brauer with extreme
minuteness of detail, neither correspond with those of the bots found by Sir John Kirk in
the stomach of the elephant shot by him in Africa, nor do they resemble the characters
presented by the bots which were procured from the stomach of one of the Indian
elephants which died in this country. The specific differences refer to size, shape,
armature, and habitat. In Pharyiigobolas the larva is, we are told, scarcely more than
twice as long as it is broad, and the two anterior rings are only a trifle narrower than
the third ring. The number and disposition of the spines are also especially note-
worthy. In Brauer's larva each armed segment displays two wddely separated circles
of spines, the circle occupying the centre of each somite consisting of large spines with
then* points directed downwards. In our larvae the armed somites at the middle of the
body support several circles of spines, which are closely set together ; moreover, whilst
the uppermost circle is composed of spines that are paramount in size, the second circle
carries rather large spines, the third, fourth, and fifth rows becoming smaller and smaller.
Besides these, it will be seen, on comparison, that there are other more or less well-
marked differences; and therefore I do not hesitate to say that the throat- and stoma ch-
bots of Elephants represent two totally distinct species of dipterous insects. — T. S. C.

SECOND SERIES. — ZOOLOGY, VOL. II. 35

256 DR. T. S. COBBOLD ON THE PAEASITES OF ELEPHANTS.

Bibliography.

Barrel, W., " Description of a new Species of Entozoon (Sclerostoma sipunculiforme) from the Intestines
of the Elephant," Proc. Zool. Soc. 1859, p. 425.

, " Description of a new Species of Sclerostoma from the Stomach of the African Elephant (Loxo-

donta africana)," Proc. Zool. Soc. 1868, p. 262.

Brauer, Dr. F., " Pharyngobolus africanus ein Oestride aus dem Rachen des afrikanischen Ele-
phanten." Vcrhandl. d. k. zool.-botan. Gesellschaft in Wien, Band xvi. S. 879 (1866.)

Cohbold, T. S.j "Description of a Species of Trematode from the Indian Elephant," Quart. Journ.
Micr. Soc, Jan. 1869; also in Supp. to 'Entozoa/ 1869, p. 80.

, " On the Destruction of Elephants by Parasites, and on the so-called Earth-eating Habits of

Elephants and Horses in India," The Veterinarian, Oct. 1875.

, " Further Remarks on Parasites from the Horse and Elephant, with a Notice of new Amphistomes

from the Ox," The Veterinarian, Nov. 1875.

, " Parasites of Pachydcrinata," in a general treatise on Parasites (Part x., sect, i., Book ii.),

London, 1879, p. 393.

Diesing, C. M., ' Systerna Helminthum/ vol. ii. 1851, pp. 176 et 560; " Revision der Myzclminthen "
(from Vienna-Academy Reports, Bd. xxxii. S. 35), 1858, S. 50; " Scchzehn Artcn von Ncmatoideen,"
Denkschr. d. math.-nat. CI. d. k. Akad. d. Wisscnsch. Bd. x. iii. 1857, S. 13, Taf. i. figs. 11-17.

Fitz, R H., "Anatomy of the Fasriola Jacksoni," reported from Boston Society of Med. Sci. in the
New York Medical Journal, Nov. 1876.

Fleming, G., "The Diseases of Elephants" (extracts from Captain Forsyth's work), The Veterinarian,
March, 1873.

Forsyth, " On the Selection and Treatment of Elephants," in his ' Highlands of Central India/ 1871.

Fiirstenberg, ' Die Kratzmilben von Menschcn und Thieren,' Leipzig, 1861 (Homopus, S. 208, quoted
by Megnin) .

Gilchrist, W., ' A Practical Memoir on the History and Treatment of the Diseases of the Elephant, with
instructions for preserving its efficiency as an animal of transport/ Calcutta, 18-11.

Jackson, J. B. S., ' Descriptive Catalogue of the Medical Improvement Society's Cabinet, Boston,U. S./
1847.

Linstow, O. von, 'Compendium der Helminthologie/ 1878, S. 45.

Me"gnin, M. P., " Sur quelques Hypopes indetermines," in Mem. sur les Hypopes, in Journ. de l'Anat.
et de la Physiol. 1874, pp. 227 et 248.

Ouchterlony, J. W., ' An Essay on the Management of the Elephant ' (without reference to parasites),
The Veterinarian, 1873.

Piaget, E., " Description d'un parasite de l'Elephant," Tijschrift voor Entomologie, 1869, p. 249.

Richter, R. C, " A new Form of Parasite (Idocoloris elephantis)" Science Gossip, 1871, pp. 131, 185,
211, and 278.

Rudolphi, ' Entozoorum Synopsis/ Berolini, 1819, p. 36.

Schmidt, M., " Die Krankheiten der Dickhauter," Deutsche Zeitsch. f. Thiermed. und vergleichende
Pathologie, vierter Bd. 1878, S. 360.

Sclater, P. L., "Note respecting Insects' Eggs on Elephants' Tusks," Proc. Zool. Soc. 1871,
p. 145.

Smith, F. W., " Diseased Elephants," (anonymously) in the Lancet for Sept. 2, 1876, p. 346.

Steel, J. H., " Fatal Helminthetic Disease of Elephants," Medical Examiner, January 17, 1878, p. 55 ;
and for Report on this outbreak as a " Case at Law," see also The Veterinarian for December 1877,
p. 886.

DR. T. S. COBBOLD ON THE PARASITES OF ELEPHANTS. 257

"White, Dr. F. Buchanan, "The New Elephant Parasite," Science Gossip, 1871, p. 234.
Williamson, T., "Natural History of the Elephants" &c, in his ' Oriental Field Sports,' London,
1807, vol. i. p. 139.

DESCRIPTION OF THE PLATES.
Plate XXIII.

Fig. 1. Head and neck of a male Sclerostoma sipuncidiforme, showing the double circle of oral denticles,

buccal cup, oesophagus, retractor muscles, upper part of the chylous intestine, and the sub-
dermal vascular network. ( x 33 diam.)
Fig. 2. Semiprofile and oblique view of the lower end of the body of a male Sclerostoma sipuncidiforme,

showing the folded hood, the points of the rays, the vascular network, and the position of the

spicules in their retracted state. ( X 27 diam.)
Fig. 3. Side view of the head and neck of Strongylus clathratus, showing two auriculate folds, the buccal

cup, and the oesophagus, (x 23 diam.)
Fig. 4. Side view of the tail of a male Strongylus clathratus, giving a complete representation of the

right lobe of the hood and its rays, and also indicating the position of the spicules and caudal

swelling in front. ( x 23 diam.)
Fig. 5. One of the spicules of Strongylus clathratus. (x 40 diam.)
Fig. 6. Lower end of the body of a female Sclerostoma sipuncidiforme. ( x 13 diam.)
Fig. 7. Another view of the lower end of the body of a female Sclerostoma sipunculiforme, showing the

altered shape of the tail when seen from the front. ( x 13 diam.)
Fig. 8. Additional side view of the tail of the male Sclerostoma sipunculiforme (to contrast with fig. 2),

showing the altered position aud direction of the spicules when exserted. The vascular

network omitted. ( x 27 diam.)
Fig. 9. Head and neck of Strongylus foliatus, seen from the front, displaying the upper and lower pair

of papillae, mouth, buccal cup, and long oesophagus. ( x 27 diam.)
Fig. 10. Profile view of the left side of the head of Strongylus foliatus, showing auriculate folds of the

skin. ( x 33 diam.)
Fig. 11. Similar representation of the right side of the head of Strongylus foliatus, with the focus altered

to show the buccal teeth. ( x 33 diam.)
Fig. 12. Left side of the lower end of the body of a male Strongylus foliatus, showing the retracted

spicules, bursal lobes, and entire system of rays on one side. ( x 60 diam.)
Fig. 13. Left side of the lower end of the body of a female Strongylus foliatus, showing the anal cleft and

subulate extremity of the tail, (x 40 diam.)
Fig. 14. Head, neck, and upper part of the body of a female Strongylus falcifer , showing especially the

outer and inner series of oral papilla?. ( x 23 diam.)

Plate XXIV.

Fig. 1. Lower end of the body of a male Strongylus falcifer, seen from the front, exposing the inte-
rior of the bursa, caudal papilla, lobes of the hood, points of the rays, and more particularly
the spicules with their sheaths and retractor muscles attached to the upper end of the blind
pouches. ( x 23 diam.)

258 DE. T. S. COBBOLD ON THE PARASITES OF ELEPHANTS.

Fig. 2. Profile view of the lower end of the body of another Strongylus falcifer, exposing the bursal
lobes of one side and the projecting spicules. ( x 27 diam.)

Fig. 3. Lower end of the body of a female Strongylus falcifer, showing the mode of termination of the
reproductive and digestive organs, and more especially the characteristic form of the tail.
( x 23 diam.)

Fig. 4. Half-profile view of the head and neck of a female Dochmius Sangeri, showing the mouth, buccal
cup, oesophagus, and upper end of the chylous intestine. ( x 33 diam.)

Fig. 5. Lower end of the body of a male Dochmius Sangeri, seen from behind. It shows the expanded
hood and entire system of rays, as well as the spicules in their retracted position. ( x 33 diam.)

Fig. 6. Profile representation of the lower end of the body of a female Dochmius Sangeri, showing the
anal cleft and finely pointed tail. ( x 33 diam.)

Fig. 7. Upper part of the body of a female Filaria Smithii, exposing the head, a double row of auriculate
folds, papilla?, oesophagus, and commencement of the chylous intestine. ( x 30 diam.)

Fig. 8. Front view of the lower end of the body of a male Filaria Smithii, displaying the spicules, the
anal opening, and the relation of the latter to the stalked papilla?. ( x 30 diam.)

Fig. 9. Profile view of the tail of another male Filaria Smithii, in which the papilla? are not seen.
( x 30 diam.)

Fig. 10. Another side view of the lower end of the body of a male Filaria Smithii. The tail is more
curved, the spiciiles being completely retracted, the papilla? of the left side being also seen
throughout their whole extent. (Flattened by pressure, and x 50 diam.)

Fig. 11. Front aspect of an Amphistoma papillatum, giving a general view of the internal organs as seen
through the transparent integuments. It also shows the oral sucker, reproductive papilla,
circular folds or ruga? of the body, and especially the disposition of the fungiform papilla?
within the concavity of the caudal sucker. ( x 18 diam.)

Fig. 12, This illustration represents a Fasciola Jacksoni, viewed from behind, and with the aid of
transmitted light. It explains the position of the oral and ventral suckers, reproductive papilla,
exserted cirrhus, two or three of the uterine coils, their contained ova, and also, more par-
ticularly, the mode of branching and distribution of the digestive tubes. ( x 10 diam.)

[ 259 ]

VIII. On the Digastric Muscle, its Modifications and Functions.
By G. E. Dobson, M.A., M.B., F.L.S.

(Plate XXV.)

Read December 15th. 1881.

IN a paper published in the ' Proceedings of the Royal Society ' for March 1881, I have
traced the origin of the oblique tendinous intersection of the Digastric Muscle to an
originally united condition of the anterior bellies of the muscles of opposite sides in
front of the hyoid bone, the superficial tendinous inscription still found traversing the
surfaces of the digastrics of many species in which the muscles are not united being
shown to be the rudiment of the origin of a tendinous band which once formed the
posterior limit and support of the united muscles. I shall now endeavour to trace the
leading modifications of this muscle and their relations to its functions.

The representative of the digastric in the lower Vertebrates, as in R,eptilia, is a bundle
of muscular fibres arising from the occiput and inserted into the posterior extremity of
the mandibular ramus, its functions being simply those of drawing the angle of the man-
dible backwards and upwards, and so separating the jaws in front. Such is its form and
such its functions also throughout the class Aves and in most of the species of Mammalia.
In many species, however, of the latter class, notably in certain Orders — the Primates and
Rodentia — we find it no longer of the simple form above described ; its anterior attach-
ment has advanced nearer to the anterior than to the posterior extremity of the man-
dibular ramus, and the muscle is found to be made up of two bellies, an anterior and a
posterior, with an intermediate tendon or tendinous intersection, thus forming a true
musculus digastricus.

In the Primates, in many species of Rodents, and in a few species of other Orders,
the intersecting tendon is connected by ligament or by tendinous fibres with the hyoid
bone, the anterior bellies of the muscles of opposite sides uniting in the middle line
across the mylo-hyoids, or connected only by the fascial expansion known as the supra-
hyoid aponeurosis which extends between their inner margins.

The leading modifications of the muscle may therefore be arranged as follows : —

I. Single (31. depressor mandibular), inserted into the posterior extremity of the

mandible, as in reptiles, birds, and many mammals.

II. Double (31. digastricus), consisting of an anterior and posterior belly, with an

intermediate tendon or tendinous intersection, inserted more or less in front into
the side of the mandible, as in the Primates, many Rodents, and some species
of other orders.
A. Not connected with the hyoid bone ; tendinous intersection oblique or
transverse.

SECOND SERIES. — ZOOLOGY, VOL. II. 3G

260 ME. G-. E. DOBSON ON THE DIGASTEIC MUSCLE.

a. With the anterior bellies united across the space between the rami of the

mandible, the mylo-hyoid muscles feeble or absent.

Ex. Gymnura rafflesii, Epomophorus macrocephalus.

b. With the anterior bellies separate, the mylo-hyoid muscles well developed.

Ex. Erinaceus ev/ropceus and other species, Cavia
apercea, Epomophorus franqueti, &c.
B. Connected by ligament or by tendinous fibres with the hyoid bone ; tendinous
intersection well developed, oblique or longitudinal.

a. With the anterior bellies united across the space between the rami of the

mandible, the mylo-hyoid muscles feeble.

Ex. Tupaia ellioti, Mm decuman/as, Ilyoxus avel-
lanarius, Cercopithecus caUithricus, &c.

b. With the anterior bellies separate, the mylo-hyoid muscles well developed.

Ex. Lor is gracilis, Homo sapiens.

In the first example cited above, Gymnura rafflesii, I found the digastric condition
of this muscle well developed. About the middle of the muscle, at its narrowest part, it
is traversed by an oblique tendinous intersection more marked internally, arising from the
upper margin, which is continued inwards and slightly forwards from its lower and
internal margin as a tendinous band across the mylo-hyoid muscle to unite in front of
the hyoid bone with the corresponding band from the opposite side. From this
tendinous raphe thus formed muscular fibres arise, which, extending forwards and
inwards, cover the anterior three fourths of the mylo-hyoid and part of the genio-hyoid
muscles, and passing above the margins of the anterior bellies of the digastric muscles of
opposite sides, are inserted with them and for some distance in front of them into the
rami of the mandible.

This horizontal muscular expansion which takes its origin thus from the tendinous
intersection and its median continuation might also be described as a deep division of
the anterior belly of the muscle, commencing at the tendinous intersection, and uniting
with its fellow of the opposite side along the middle line between the jaws. Its posterior
margin, the tendinous raphe above described, is concave, quite free from the hyoid bone,
but attached to a fascial aponeurosis, which passes backwards over it and the sterno-
hyoid muscles. On removing this muscular expansion the mylo-hyoid muscles are found
beneath, very thin, not extending half the distance between the hyoid bone and the
symphysis menti.

In Epomophorus macrocephalus, minor, labiatus, and gambianus the anterior bellies
of the digastric are double, and the deep lamina> united not only across the space
between the rami of the mandible, as in Gymnura rafflesii, but their union extends so
far backwards that when the head is bent forwards then' posterior free margin passes
behind the hyoid bone. The mylo-hyoid muscles are completely absent, their functions
being evidently performed altogether by these greatly developed digastrics. It is most
remarkable, as showing what extraordinary structural differences may occur in the
anatomy of species to judge from external characters very closely allied, that in
Epomophorus franqueti, comptus, and pusillus the digastrics arc single, and not united

MR. G. E. DOBSON ON THE DIGASTRIC MUSCLE. 261

in front, and the mylo-hyoid muscles are well developed ; nevertheless, the presence of a
nearly vertically transverse tendinous inscription across the digastrics of those last-
named species, corresponding precisely iu position and direction to the origin of the
tendinous raphe bounding the united anterior bellies of the muscles behind in the first
named, indicates a united state of the digastrics in a common ancestor. It is espe-
cially interesting to note that in these species, where union of the muscles extends so far
backwards, the tendinous intersection is vertical or nearly so, and the direction of the
fibres of the deep or internal lamina transverse, while those of the superficial external
part spring from it at right angles to its direction and pass forwards to their insertion
into the mandible. Here, then, we have a vertical tendinous intersection unconnected
with the hyoid bone !

In the genus Erinacetis, which with Gymnura makes up the family Erinaceidae,
although the digastrics are no longer united, a superficial oblique tendinous inscription
occupies precisely the same position as in Gymnura, though much less developed ; it
extends from the upper margin of the muscle (which is nearly of the same calibre
throughout), and on reaching its lower margin is continued inwards into a fascial
expansion extending between the anterior bellies of the muscles of opposite sides,
covering the mylo-hyoid muscle, but no muscular fibres arise from it as in Gymnura;
nevertheless this fascial expansion evidently represents the muscular sheath of the
expanded and united digastrics iu the ancestral forms.

In Centetes ecaudatus and semispinosus, in Pteropus eclulis and medius, and probably
in all the species of that genus, in Epomophorus franquetl, comptus, and pusillus, in
Megaderma lyra, in Cavia apercea, in JPkoca communis, and in other species, though the
fascial expansion be absent or feebly marked, the transverse tendinous inscription is
traceable, although its presence in some, as in Centetes ecaudatus, is indicated by a faint
superficial oblique line surrounding the muscle.

In Tupaia ellioti the anterior bellies of the digastrics also unite in the middle line
between the jaws; but there is no separate superficial external lamina in direct con-
tinuation with the posterior belly, as in Gymnura rafflesih. The intersecting tendon is
narrower but more distinct, arises, as in that species, from the upper margin of the
muscle, and is continued downwards, forwards, and inwards to unite with the correspond-
ing tendon from the opposite side across the mylo-hyoid, precisely as in Gymnura
rafflesu ; but unlike its free condition in that species it is closely adherent by its deep
surface to the mylo-hyoid muscle, and connected with the hyoid bone by a strong fibrous
aponeurosis. The united anterior bellies pass forwards, taking their origin from the
tendinous raphe, and, separating slightly near the symphysis menti, are inserted into the
rami of the mandible, and by a fibrous aponeurosis into the symphysial angle. The
mylo-hyoid muscles, nearly wholly concealed by them, are feebly developed.

In If us decumanus and in other species of the genus the digastrics are united in front,
the posterior margin of the united muscles being formed by the united intermediate
tendons which are connected with the hyoid bone ; in Myoxus avellanarius the inter-
mediate tendons are even better developed, and the united anterior bellies are closely
connected with the mylo-hyoid muscles. In Cercopithecus callitrichus, in Troglodytes

262 ME. G. E. DOBSON ON THE DIGASTEIC MUSCLE.

niger, and in many other species of Primates, the anterior bellies of the digastrics are
united, and the intermediate tendon is well developed and connected with the hyoid
bone.

In Man, as an anomaly, the fibres of the two anterior bellies have been found blended
together, or the anterior belly has been found double, as in Gymnura, the deeper part
uniting with its fellow in a median raphe, covering the mylo-hyoid muscle, or the inter-
secting tendon is not unfrequently continued from one side across to the other, forming
a zone immediately above the hyoid bone, to the body of which it is tied down by fibrous
tissue ; from the upper surface of this the anterior bellies arise.

This united condition of the digastrics in man is very interesting, as it indicates rever-
sion to a condition common in lower species of Primates.

If now we arrange the species of Mammals referred to above according to the relations
of the digastric to the hyoid bone, we shall find that they fall into two groups : — 'in one
the muscle is connected with the hyoid, and the species swallow their food while in the
erect position, with the head bent forwards upon the chest and the long axis of the
cavity of the mouth at right angles with the oesophagus ; in the other this muscle is free,
and all the species feed while resting on their anterior extremities, having the long axis
of the mouth in a line with the oesophagus. In the latter, where a tendinous intersection
exists, it is, as I have shown, either the origin of a tendinous band, as in Gymnura raf-
flesii, which, passing obliquely inwards and forwards across the mylo-hyoid, unites with
its fellow of the opposite side, or it is very small, forming a mere superficial tendinous
inscription, as in Erinaceits, indicating a united condition of the muscle in ancestral
forms.

Thus among certain Rodents and arboreal Insectivora, as the species of Tvpaia, which
habitually sit erect when feeding, holding their food between their fore feet, the anterior
bellies of the digastrics are large and united, and the intermediate tendons well developed
and connected by fascial bands with the hyoid bone and by their deep surfaces with the
mylo-hyoid muscles, as may be seen in the species of Muridce and Ifyoxidce, of which ex-
cellent examples are readily available in the rat and common dormouse. In the water-
vole (Arvicola amphibitts), however, the digastrics are connected together in front by fascia
alone, and the upper margin only of their intermediate part is tendinous and not con-
nected with the hyoid bone. These animals live on vegetable substances obtained while
swimming, and habitually hold the head stretched out in a line with the body. Again,
an instructive example is met with in the great rodent moles (Bathyerges marltimus)
of South Africa, and in the species of Spalax and Siphneus, which feed chiefly on the
roots of grasses, and while feeding necessarily hold the head in a line with the body,
or raised upwards ; in them, accordingly, the digastric is quite simple, without tendinous
intersection of any kind, and unconnected with the hyoid bone, a true depressor niandibulce.

In those animals, then, which habitually swallow their food while the head is bent
forwards at a right angle with the neck, the digastric has functions other than those of
a simple depressor mandibuloe ; it is an important agent in the act of deglutition, such, in
fact, as we find it in man. Where, as in reptiles, birds, and most mammals, the position
of the mouth with respect to the oesophagus during the act of swallowing the food is

ME, G. E. DOBSON ON THE DIGASTEIC MUSCLE. 263

almost in the same right line, deglutition is easily effected hy the mylo- and genio-
hyoid muscles drawing the hyoid hone and larynx forwards and upwards, so as to allow
the masticated mass to get hehind them, and so hring it within the grasp of the pharyn-
geal muscles ; hut in those animals which feed while in the erect or semi-erect position
and the head is hent forwards, so that the cavity of the mouth is at right angles with
the oesophagus, it is evident that deglutition must be a much more complex action. In
that position the mylo- and genio-hyoid muscles are relaxed, and cannot act efficiently
in drawing the hyoid hone upwards and forwards, so as to allow the masticated mass to
pass into the oesophagus, into which it has to pass, in fact, round an angle. The difficulty
is got over by the connection of the digastric with the hyoid bone. This muscle during
the act of deglutition causes the hyoid bone, larynx, and base of the tongue to move
through a segment of a circle, the anterior part of the muscle drawing these parts
forwards ; they are then elevated by the joint action of the anterior and posterior bellies,
and finally drawn backwards by the posterior bellies, so as to force the masticated mass
into the oesophagus.

This action of the digastric is well known, but its relation to the position of the head
while the animal is feeding appears not to have been generally recognized.

With respect to the great development and united condition of the anterior bellies of
the digastrics so general in Quadrumana, and not uncommon in Podentia, it appears to
me to be related to the inefficiency of the mylo- and genio-hyoid muscles in the habitually
bent forward state of the head. Their separate condition in man is probably due to
shortening of the rami of the mandible and greater depth of the symphysis menti, which
bring the mylo- and genio-hyoid muscles into a more favourable position for action, as
well as affording more extensive attachments. Nevertheless united digastrics have been
noted as a muscular abnormality in a few cases, evidently recurrences of a once normal
condition of these muscles in ancestors provided with much longer jaws than we have ;
and I suspect that, when the muscular anatomy of the prognathous inhabitants of Africa
has been more fully examined, united digastrics will be regarded as a much less uncommon
human muscular anomaly than it uow is.

EXPLANATION OF PLATE XXV.

Fig. 1. Myoxus arellanarius. United digastric muscles (enlarged).

Fig. 2. Mus natalensis. Dissection of inferior gular region, showing the position and form of the

united digastrics.
Fig. 3. Tupaia ettioti. Dissection similar to the preceding.
Fig. 4. Ditto. Digastric muscle, showing the oblique tendinous intersection passing across from the

infero-esternal margin of the posterior belly of each moiety to support the posterior free margin

of their united anterior bellies, uniting with the corresponding tendinous band from the

other side (enlarged) .
SECOND SERIES. — ZOOLOGY, VOL. II. 37

264 MR. G. E. DOBSON ON THE DIGASTRIC MUSCLE.

Fi<*. 5. Gymnura rafflesii. United digastrics, showing conditions similar to those described in fig. 4.
Fig. 6. Erinaceus europaus. Single digastric muscle, showing an oblique tendinous intersection in a

position similar to that in Tupaia and Gymnura, but rudimentary and unconnected with that

of the corresponding muscle of the opposite side.
Pig. 7. Epomophorus franqueti. Dissection similar to that shown in fig. 2, showing the position and

direction of the muscular fibres (see text).
Fig. 8. Cercopithecus calRtrichus. Anterior bellies united, and intermediate tendons well developed,
tig. 9. Troglodytes niger. Dissection of the infero-lateral side of the neck, showing the positions and

attachments of the united digastrics (after Vrolik).
Fig. 10. Homo sapiens. Dissection showing an abnormal united condition of the digastric homologous

to that in fig. 8.

[ 205 j

IX. On the Clasping -org cms ancillary to Generation in certain Groups of the Lepido-
ptera*. By Philip Henry Gosse, F.B.S. [Communicated by E,. McLachlan,
F.B.S., F.L.S.)

(Plates XXVI.-XXXIII.)

Head May 4th, 1882.

Introductory Bemarks : Existing Authorities.

1 HE lovely insects, of which I am about to treat, have been the objects of so much
ardent admiration and intelligent curiosity, so long have they been collected from all
regions, so many cabinets are filled with them, so many treatises have been written and
compiled about them, that one may well be accused of presumption in daring to
surest the existence of an uncultivated field of observation in their structure. Yet
I venture to lay before the Linnean Society a series of facts, yielded to my own examina-
tion in the study of these exquisite creatures, which, while they elicit interest and
admiration in an unusual degree, seem to me to have beeu hitherto almost unsuspected.
I refer to the complex apparatus with which the male Butterfly has been furnished for
the prehension of the female during the copulative function.

In most of the families of the Bhopalocera the organs devoted to this purpose are more
or less patent when sufficient magnifying power is used; and these are often exceedingly
curious and exceedingly various. If the Papilionkhc have seemed less profusely supplied,
it is, doubtless, because in them the prehensile apparatus is concealed beneath special
organs (calces), which do not to the eye present any feature differing from the scale-
clad body. The family, moreover, is almost wholly Trans-European ; and thus, though
very conspicuous, they lie somewhat beyond the range of ordinary scientific inves-
tigation.

I hope I am doing no injustice to my predecessors, in assuming these researches to be
new. The older authors have little more tban vague allusions to the existence of
auxiliary prehensors, and nothing, that I am aware of, concerning those of the Fquites.
Herold (1815) describes and figures f the "spoon-like" valves, and the "triangle-
piece " in Bieris Brassicce. Burmeister (1832) describes J the outer and inner valves in a
Hawk Moth. Kirby and Spence (1828) devote § a paragraph to the prehensors of the
class, but dismiss the Butterflies in a few lines. Siebold and Stannius (1848) assign ||
to the prehensors of the Lepidoptera one fourth of one sentence.

The most important contribution to the subject seems to be a memoir f read before
this Society by Dr. P. Buchanan White, December 21, 1870. This elaborate and very
valuable article, illustrated by more than 220 figures, might well seem to leave little
room for my researches, but for two reasons : — 1, he limits his studies to the Butterflies

* A preliminary sketch of this article will be found In the Royal Society Proceedings, xxxiii. (18S1), pp. 23-2".
t Entwickel. p. xiii. pi. iv. t Manual, p. 216 (Eng. cd.).

§ Intv. to Ent. iv. p. 579. 1 Anat. Inv. p. 462 (Amer. ed. |.

*il " On the Male Genital Armature in the European Rhapalocera," Trans. Linn. ISoe. 2nd ser. Zool. i. p. ;:"7.
SECOND SERIES. — ZOOLOGY, VOL. II. oii

260 MR. P. H. GOSSE ON THE CLASPING-ORGANS

of Europe, of which only three species come into my subject ; 2, he does not look within
the anal valves, within which my study begins.

Somewhat earlier than this, however, viz. in the ' Proceedings of the Boston (U. S.)
Soc. of Nat. Hist.' for April, 1870, Messrs. S. II. Scudder and E. Burgess had published
a conjoint memoir, of much interest and value, and illustrated by many highly magnified
figures, "On Asymmetry in the Appendages ... in the Lepidopterous Genus
msoniadesr The specific variety and individual uniformity, which I hope to show in the
genital armature of Papilio, mark the corresponding organs in this genus of Hes-
periadre; nor is the apparatus less elaborate or less curious. The learned authors
designate by the terms " clasps " and « upper organ," what Dr. White calls " harpagines"

and " tegumen."

Prof. Grabor, in his valuable work, ' Die Insekten ' (in < Die Naturkrafte'), 1877, has
devoted half a 'dozen pages to the male genital auxiliaries (ausscrc Hilfsorgane des Ilin-
terleibes) ; but has taken no illustration from the Lcpidoptera.

I am indebted" to my honoured friend Prof. Westwood for bringing to my knowledge
some descriptions and figures by Dr. Do Haan on the subject, which, so far as they go,
do really anticipate my own. The book is rare and difficult to consult; but my son,
Edmuncl William Gosse, has made for me, from the copy in the library of the British
Museum, a translation of the Dutch test, and careful tracings of all the figures that

relate to the subject.

It is a thin folio, a livraison of a large publication, entitled ' Verhandehngcn over de
natuurlijke gesehiedenis der Nederlandschc overzecsche bezittingen (Transactions re-
garding the°]STat. Hist, of the Dutch Transmarine Possessions), by various authors.
The sub-title of the treatise itself is " Bijdragen tot de kennis des Papilionidea : door
Willem de Haan." Leiden, 1812.

The observations of this author on the genitalia, since they are brief, and the work is
little available to English students, I shall be excused for giving in externa.

"The Sexual Organs.
« These parts are especially in the male, very various in form. The outermost valves \_klepperi] of Oml-
thoptera Amphrisius shut closely against each other, and conceal two lateral appendages \_zij deling scfa
aanhangsels = the harpes of the. following memoir?] , which turn over at the end in the form ot a hook, and
are provided with spines along the upper edge. The truncated, and sometimes even hollow, end ot the
hinder part of the body, which lies within these valves, bears upon the upper edge a pointed spine bent
downwards r=miCus, P. II. G.] , with two lateral plates \_zijdelings plaatjes, = scaphium, P. H.G.?] , which
lie close to it. Prom the middle of the hinder part of the body the penis [rocde] is exposed, which is
bristly [kraakbeenig'], straight, blunt, and shorter than the lateral valves.

' " Those of Papilio Memnon, Pammon, Helenus, Machaon, are similar to this, except that the lateral
appenda-es lie along the lower edge, and the hook on the back is usually more elongated; m Memnon
the penis is thrown over from below. In P. Coon the outer valves are wholly opened from above, and,
to a great extent, also from below; so that the inner parts are scarcely protected by them : the penis is
very pointed and a little longer than the valves. P. Liris and Pohjdorus have the hinder part of the
body naked ■ the valves arc very short, armed from within to the top with a hook, and the lateral
appendages (of Amphrisius) arc placed under it; the spine of the back is blunt; besides this there arc
also two spines present on the upper edge, and these arc curved inwards; so that altogether the outer

IN CERTAIN LEPIDOPTEEA. 267

edge is provided with seven appendages, which all rise free from one another. Between these lies the
penis, which, as if pressed together, has a sharp edge above and below.

"In Thais Hypsipyle the valves are turned towards one another at the end, open above and below,
toothed along the upper edge, and armed along the under edge with a sharp spine; the lateral appendages,,
and the spine at the back, arc very small, and, indeed, scarcely visible. Parnassius Apo/lohas the valves
puffed out, and gaping wide open ; the lateral appendages are entirely free, of the same length [eens zoo
/any], and in the form of two half-sheaths; the back-spine is broad, and ends in two short points.
Eurycus Cressida has two flat valves placed on the under side of the body, which are covered from above
by the last two abdomen-segments, and show on their inner edge, close to the joint, an indentation.
The penis is pointed, and longer than the valves. Doritis Apollina has two little pointed valves, also
placed under the last two segments."

The work is illustrated by eighteen figures of: the sexual parts of eight species, on a
scale considerably smaller than mine. They are thus distributed : —

Ornithoptera Amphrisius. . . 3 figs. Thais Hypsipyle 2 figs.

Papilio Coon 3 ,, Parnassius Apollo 2 ,,

Liris 2 ,, Doritis Apollina 2 ,,

Leptocircus Curius 2 ,, Eurycus Cressida 2 „

It so happens that not a single species illustrated by Dr. De Haan is identical with
any one of mine ; though two approach very nigh, viz. his 0. Amphrisius to 0. Bha-
damanthus*, and P. Coon to P. Boubledayi.

My own attention was first directed to the subject when, examining the anatomy of
the remarkable genus Euryades, I referred for comparison to the structure of the genital
organs in Papilio and Ornithoptera. The elaborate apparatus which I found hidden
beneath the valves, like delicate surgical instruments in their protecting cases, drew me
on into the examination of one species after another, until I had studied and figured
nearly a hundred members of these two genera.

Dr. Buchanan White says t : — " I regret very much that, from Avant of material, I have
been unable to examine, amongst the llhopalocera, any but European species. It is
much to be desired that some one, who has at his command a large collection of the
butterflies of all regions, should investigate, more extensively than I have been able to do,
the structure of the genital armature."

Nothing, unless it be the exquisite beauty of the workmanship, is so astounding as
the variety, in form and detail, presented by these hidden instruments. Out of the
number that I have examined, 1 have not found any two species whose apparatus is
alike, or even so nearly alike that a moment's observation is not sufficient to show the
difference. I do not think this rule of diversity holds in respect of other important
organs. I do not know that microscopic examination would show that the antenna?, or
the tarsi, or the spiracles, vary markedly in every species % .

* Perhaps his Amphrisius is the species which in M. Lucas's ' Lepid. Exot.' is so named in the plate, but lihadu-
nanthns in the text. It is certainly not the AmpJirysus of Cramer's plate 219.

t <>i>. cit. p. 366.

$ " Aueune autre partie du corps des papillons ne denote des variations si nombrouses que la terminalo avec les
organcs nenitaux des deux sexes; variabilite qui ne nous permet pas de nous occuper de ce sujet d'une facon plus
ge'nerule, parce que son explication detaillec est l'objet d'une monographic spe'ciale." — De. Bukmeister, Lepid. d. 1.
B«p. Anj. p. 21.

38*

268 ME. P. II. GOSSE ON THE CLASPING-OEGANS

It might seem that, by the aid of organs so uniformly present, so easily examined, and
so varied in different species, while constant in the same, great facilities must he
afforded for the determination of specific identity and limitation. Yet, in practice, I fear
this cannot he carried out, without severing species which otherwise seem most closely
allied, and linking others which have little else in common. Look, for instance, at the
three African species Papilio Promhts, P. Nireus, and P. Pkorcas; how consimilar are
these in their forms, colours, and markings ! yet how diverse in their whole prehensile
apparatus ! The shape of the valve, its fringing ; the shape of the harpe, its armature ;
the uncus ; the teeth of the scaphium ; and, finally, the penis ; — all these differ signally
in one from their conditions in the others, as may he seen, at a glance, from the following
pages. The like terms might he employed concerning P. Demoleus and P. Erithonius.

It needs scarcely to he told that all the following observations were made on dried
specimens. The desiccation of the soft tissues, causing them to shrink, throws into
distinct prominence the hard chitinous organs with which these pages have mainly to
do ; and their form is not affected by drying. Indeed, the density and unyielding hard-
ness of this material are manifest by the depth of colour it can assume, often approaching
to black, by its brilliant polish of surface, by its transparency like that of glass, and
by the delicacy with which it is fashioned into the thinnest edges and cut into the
sharpest teeth, which, strange to say, we never see blunted by use.

Let it be remembered that, while a considerable number of the following descriptions
and figures have been confirmed by observations repeated upon two, three, or even more
successive examples, many (somewhat more than half) rest upon individuals. These,
though made with all care, are like the «?rn^ Xeyofieva of the critics, and must be accepted
with a certain measure of caution.

Manipulation.

My methods of manipulation are of the simplest. With a penknife, worn by age to a
very fine and keen point, I make incisions, vertically as the insect lies on its side before
me, along the hinge-line of one valve (I have usually selected the right) : presently, using,
with a gentle violence, such a lever as the tip of a toothpick, the valve is prized off. By
practice, I am able to do this with very little chance of injury to the specimen ; and when
my examination is completed, I restore the valve to its place, with the minutest touch of
gum-tragacanth ; so that, when it is dry, and the butterfly returned to the cabinet, I am
often quite unable, some lime after, even myself to determine, with certainly, of two or
three examples in the drawer, which it was that I had nsed for the operation*.

The detached valve is then submitted to strict search with the various powers of a
triple lens, reaching in combination to 24 diam. When the information deducible from '
this aspect seems exhausted, if the harpe appear to promise any more, I essay to lift it

* The circumstance thai the exotic Papiliones are often of high monetary value has no doubt much impeded omj
acquaintance with their anatomy. Our museums and private collections are rich in species; hut few possessors are
willing to submit their treasured beauties to the anatomist's scalpel. Perhaps Prof. Siebold's sarcasm is not yet
wholly without force : — "Mosi collectors seem ambitious only to keep their butterflies neat and untouched, or to
gain thereby advantages other than scientific" (--On the :, of Parn. Apollo" Stett. Ent. Zeit.. 185] |.

IN CERTAIN LEPIDOPTEEA. 269

from its adhesion to the valve-cavity with the same knife-point. I do not always suc-
ceed ; but if I do, I have an object, often of brilliant transparency and lustre, and of
extreme delicacy, which I can put on a slip of glass, and transfer to the stage of a
compound microscope, using any powers that I please.

The organs which project from the abdominal cavity cannot be detached with certainty
of success. These, therefore, I usually view with the; lens only*. After exhaustive study
of these parts, dry, since the scaphium and, in part, the penis are composed of soft tissues
which change form in drying, I introduce into the cavity, with the tip of a fine hair-pencil,
a drop of clear water. This is presently absorbed ; the superfluous water is removed
with a point of blotting-paper, and the organs are again examined as before. The shri-
velled parts have now become much more plump, assuming something approaching the
form which they had daring life. The study of them in this condition often reveals facts
that had been obscure before.

On detaching a valve there are often found many plume-scales and atoms of meconium,
scattered over the organs within, as well as all about, the anal cavity. Since these
hinder distinct vision, I remove them by touching the parts with the tip of a fine camel's-
hair pencil slightly moistened, wiping the tip on a linen cloth after every contact.

Curious records of past history may be read in such impediments. Minute fragments
of a pulverulent, drab-coloured, chalky substance are, as I have observed, occasionally
found adhering to the harpes. On one occasion I found, on carefully removing one of
the valves from a cabinet-specimen of Omithoptera ffliadamanthus <3 , the cavity quite
full of this substance, partly in coarse powder, and partly in somewhat coherent lumps
of the size of mustard-seed downwards. Now, I have no doubt that this male had
effected coitus with a female, at so early a period after her evolution from pupa, that she
had not discharged the faecal accumulation of the pupa stage, which subsists " sous la
forme d'un fluide jaune ou bruu, comparable an meconium des enfants nouveau-iiL'S "
(Burmeister). The excitement of the sexual copula would provoke the instant discharge
of the meconium, a large portion of which would fill the valvular cavity of the graspin"-
male; and this, presently coagulating, remained to be at leugth revealed by my dissect-
ing scalpel.

A phenomenon exactly similar occurred with a specimen of JPapilio Zalmoxis. But in
this case the cavity was occupied with a dirty mingled mass of meconium and body-scales ;
and that so fully as completely to conceal all the organs, till it was gradually extracted.
These scales were surely those of a female abdomen, removed in coilu: the meconium
discharged at the same time had agglutinated the scales into a composite mass, which,
in drying, had divided into fragments and coarse powder; and these, the male having
been presently captured, had remained a record of the history.

See also, for similar conditions, the account of Papilio Action, and of not a few other
species.

The illustrative drawings have all been made, not with the aid of the camera, but by
the eye and hand alone. They have not been drawn to scale ; but those of the valves

* I have occasionally viewed these as opaque objects under condensed reflected light both artificial and solar;
but the result has been (in general, though with exceptions) hardly worth the pains.

270 ME. P. II. GOSSE OX THE CLASPLNG-OEGANS

follow a pretty uniform ratio, which may be set down as about 7 diain., with which ratio
the figures of the other organs may be compared.

A General Description of the Organs.

The organs with which I propose to deal, as limited to the male sex of Butterflies
belonging to the two genera Ornithoptera and Papilio, are the following: — 1. The
Valves; 2. the Harpes ; 3. the Uncus ; i. theScaphium; 5. the Penis.

1. The redoes.

To the overlapping edges of the eighth abdominal segment, from the indefinite points
where the projecting angular portion produced by the rami of the uncus merge into the
vertical, to two prominent rounded pieces Avithin the ventral margin, there are united,
by free-working joints, two broad plates, well known under the name of valves *. Every
collector is familiar with these organs ; for he has recourse to them to determine, at a
glance, the sex of his specimens, since, in the grand genus Pap'dio, they arc peculiar to
the male.

Their outline is usually more or less trigonal, often becoming semiovate, or even sub-
circular. To be more precise, the prevailing shape is that of an oval, or rounded rhom-
boid, cut off obliquely, so that as we look on the inside of the right valve the tip points
to the left — that is, toward the dorsal side.

They are clothed on the exterior with scales like those of the body. Their direction
follows the general plane of the sides of the abdomen ; they are convex externally, con-
cave internally. Their free edges are commonly furnished with a fringe of modified
scales, taking the form of thick hairs, which are parallel, close-set, often dense, and often
long ; thus these practically enlarge the area of the valves, and help to exclude, more
effectually, extraneous matters from the genital cavity when the valves are closed.
Sometimes dense brushes of long hair, of quite different structure, arise from within the
interior of the valve, which spreading, conceal the margin, and greatly modify the
apparent outline ; e. g. JPoliccnes, Ucaleyon.

The friuged edges come into contact when the valves are at rest, inclosing a spacious
chamber, the anal, or genital, cavity. The edge is occasionally turned over, narrowly,
within ; e. g. Ornithoptera HaUphron ; and still more commonly the interior surface
continues fiat and level for some distance within the margin, then abruptly drops to the
central cavity; e. g. Agamemnon, Hesperus, Lycidas, Anchisiade-% Eurylcon. Occa-
sionally a sort of wall, of distinct form and surface, borders the edge, and overhangs the
interior ; e. g. Zalmoxis. The floor of the interior, whether fiat or hollow, is often
studded with short scattered hairs, each springing from a dimple, or crowning a minute
round hillock, where it is not occupied by the harpe : but sometimes this surface is
cloth-like; e. g. Erithonius, Lycidas.

* A better name would be tegumina, covers; but this word has been chosen by Dr. B. White for a very different
neighbouring organ. Colei (icoXebs, a sheath or scabbard) might be used ; or, to avoid ambiguity with the genus
Culms of the botanists, it might be spelled in its Ionic form, Kovkebs. But I adopt the appellation already in
familiar use.

IN CERTAIN LEPIDOPTERA. 271

In those abnormal cases in which the valves arc partly or wholly aborted, e. g.
Diphilus, Polydorus, Hector, the thatch of long hair that occasionally projects from the
dorsal arch of the eighth segment, and thus fills up the triangular space left by the
receding dorsal edges of the valves, is particularly abundant.

The extremity of the valve shows, in almost all cases, some indication of a projection
of the outline into a point. Usually this part is as soft and flexible as any other ; but
in some instances, c. g. Scmperi, Ucalegon, there seems, in its definite form and darker
hue, an approach to that condition which occurs in other families of the Rhopalocera,
in which the valve and the harpe become one organ *.

Mr. Wallace, in his very valuable Memoir " On Variation, &c, in the Papilionidae of
the Malayan Region " f, is inclined to rest on the peculiarities of the valves of Ornitho-
ptera for generic diagnosis. Having rejected certain points that have been relied on, he
says, " There remain, therefore, only the characters of the perfect insect, the most
important of which are the anal valves in the male. These are very large, ovate or
rounded, coriaceous, and not hairy, and are furnished with projecting points or spines
(sometimes very conspicuous) which serve to attach the male more firmly to the female
in copula. In several species, I have observed, these points or hooks were buried in the
protruded anal gland of the female, and thus effectually prevented the great weight of
the insects causing them to separate upon suddenly taking flight " 1.

2. The Ilarpcs.
The interior surface of each valve, which, as I have said, is hollow, is the seat of a
peculiar organ, which appears to take a very prominent part in the prehensile function.
The floor proper of the valve-cavity is a thin membrane of slight tenacity, more or less
of a dark brown hue, and of a dull unreflecting surface. Within the hollow lies a plate
of what I presume to be pure chitine §, usually as transparent as glass, but tinged with
a yellow-brown hue, thickening at various parts, especially at its margins and irregular

* Dr. P. B. White's memoir affords abundant examples of this combined condition. The transition seems to be
made in the family (or subfamily) Pieridoe. (See the supplementary note to this Memoir.)

t Trans. Linn. Soc. xxv. p. 35, 1804.

% But the valves in Orniilioptera Arruana and Amphrysus are densely fringed with hair-scales : Papilio Semp ri
has as distinct a finger-point to the valve as 0. Amphrysus : while the broad rounded outline is seen in the valves of
P. Zalmoans, Axion, and Broinius. This projecting finger-point, moreover, while well marked in 0. Amphrysus and
Bhadamanthus, becomes feeble in Arruana and Haliphron, evanescent in Remus, and is absolutely wanting in
Brookeana, where the valve is surrounded by a fringe of unusual density. But generic lines arc not mathematical
lines: and the features mentioned by Mr. Wallace are certainly characteristic, and, with the prevailing form of the.
harpe, go far to establish Orniiliopti ra as a good genus.

Mr. Wallace's observation about the function of the points is most interesting ; would that we bad more of such !
but, if I may venture to correct the dictum of so great an authority, I would suggest that, most probably, it was
not the flexible finger-point of the valve that had seized the protruded vulva of the female, since it has no resistent
power, but the harpe, well toothed, and of almost metallic hardness, that was concealed just beneath the valve-tip.
Yet, in some cases, as in P. Vcaler/on, I am not sure that the valve-point may not have a harpe-power — as I presume
that it has in some Pieridffi, as G'aZUdryas.

§ The various degrees of flexibility found in the integuments of insects i.- believed to depend mi the extent to
which they are permeated by this substance, " a peculiar azotic matter insoluble in caustic potass " (Siebold, 'Anat.
Invertebr.' Loud. 1854, p. 401).

272 MR. P. H. GOSSE ON THE CLASPING- ORGANS

ridges, where the substance deepens in hue, often to an intensity almost black (but still
like black glass), and where it manifestly acquires great density, elasticity, and hard-
ness, with a reflective power equal to that of polished glass. The gleam and glitter of
the surface imparts a wonderful attractiveness to these most diversified organs. The
parts thus thickened are also elevated, not merely in the bounding walls and ridges that
I have mentioned, but, in general, by the separating of a portion of the organ from the
cavity-floor ; so that this particular part shall be elevated, and projected freely into the
inclosed space between the valves. And these projected portions either take the forms
of curved acute spines, or, more frequently still, are notched into sharp teeth standing
in serried rows.

To attempt to indicate the various forms which tbese curious weapons assume, would
be to anticipate the greater part of the descriptive pages which are to follow; for the
variety is about commensurate with the number of species examined. I propose pre-
sently to attempt a rough casting of them into groups ; but it will be little more than
an arbitrary classification, useful as an aid to memory, rather than any thing else.
Projecting claws, hooks, pikes, swords, knobs, and strange combinations of these, abound,
whose prominences are very commonly brought to a keen edge, and then cut into sharp
teeth. The species of one group, Agesllaiis, Parmatus, have harpes in which many
such projections arc strangely crowded. In others, Ulysses, Macedon, the formidable
teeth are themselves cut into many minuter toothlets of wonderful delicacy.

Sometimes this polished weapon occupies almost the whole area of the valve-cavity,
Agesilaus, Podalirius, Erithonius ; sometimes it is a slender staff or rod, Memnon,
Erechtheus, Polymnestor, with the extremity expanded to form the weapon, whether
axe, or sword, or saw, or spear-point. Whatever the form, the base is always expanded,
often with ridges (like as where the trunk of a tree dilates into roots), spreading over the
basal edge of the valve, and manifestly stretching beyond its area, as is often very clearly
seen when the valve is forcibly detached. Examples of what I mean may be seen in
my figures of Papilio, Axion, Ascalaphus, Pammon, Ilacedon, Childrence, Verlumnus,
Zalmoxis, and several others. In all these a kind of disk is observed ; which more or less
clearly shows that there had been a dilated attachment of the harpe to a surface which
had lain in a plane quite different from that of the valve, face to face with it, indeed,
close to its hinge. (See, specially, Euryleon.)

Hence I very confidently infer the function of two short thick eminences of hard
chitine, having convex surfaces, which are generally to be observed occupying the bottom
of the abdominal cavity, when the valves are removed ; e. g. 0. licmtts, P. Pammon,
Mlietenor, Mayo. They afford, doubtless, a broad and firm basis for the attachment of
muscles which move these important prehensores, so constantly exposed to shocks and
violent strains. And the fact of such an attachment explains an otherwise puzzling
phenomenon which not very infrequently occurs, — that, when the valve is detached by
force, it occasionally comes away empty, the chitinous armature being left behind. (See
infra, page 298.)

If there existed any reasonable doubt of the work performed by this well-armed organ,
it would be removed by a due consideration of that curious accumulation of foreign

IN CEETAIN LEPIDOPTEEA. 273

matter to which I have already alluded, — the thick mass of dislodged hody-scales with
which we frequently find the chitiuous armature clogged, particularly its serrate parts.
How came those scales there ? The answer is patent. This apparatus has been provided
to enable the male butterfly to grasp and hold the female during the process of impre-
gnation. And these accumulated scales show that this very individual insect had been
so engaged only just before its capture and death.

I have hence no hesitation in assigning a distinctive epithet to the organ in question ;
and it is known, throughout this memoir, by the term Harpe (apirri, a grappling iron)*.

It is the rule, subject to very rare exceptions, that when the two valves are closed, as
generally during life, the chief armed parts of the two harpes approach each other at a
poiut in space, within the post-anal cavity, exactly where the tip of the uncus comes
down from above. Hence I infer that the function of this latter organ is similar and
ancillary to that of the harpes — the three combined constituting a threefold grip, which
it would be difficult to break.

It seems scarcely credible that elaborate instruments, such as these, so constantly
present in this popular family, and so readily found, should not have already been the
subjects of ample examination and familiar knowledge. Yet, with the exception of
Dr. De Haan's brief allusion to the " lateral appendages," and one equally brief in Dr.
Burmeister's latest work, the ' Lepidoptera of the Argentine Republic ' t, I know not any.

I have already spoken with admiration of the surprising variety which the prehensile
apparatus displays ; it is in the harpes that the extreme diversity mainly resides.

Yet it is by no means a vague, come-by-chance variation ; it does not extend to
individuals of the same species. I have, in many instances, examined several examples
of the same species, and have always found that the identity of the harpe is wonderfully
close and minute, the diversity very trifling.

3. The Uncus.

The eighth segment of the abdomen, in the male of Ornithoptera and Papilio, in
general, has the posterior % outline of its dorsal arch produced in the middle line, and
terminating in a point, with receding sides ; so that, looked at vertically from above, it is

* Dr. WTiite has used the term harpago for the organ which, in the other Bhopaloeera, appears to represent the
valve and harpe united. But, in the Papilionidae, where these are separate, it is de irahle that they should receive
separate designations. The terms harpago and harpe are sufficiently distinct : while they bear a relation to-each other
not unlike that of the things designated.

t " Entre ces deux valvules, se trouve, dans lo fond de la cavite, l'ouverture sexuelle, aftcompagne'e, principalement
chez les males, par des appendices en forme de tenaille, qui renferine 1'organe male, le penis." I conjecture that, by
these " appendages in form of pincers,'* he means the harpes. But the term is vague, and T do not feel quite sure.

J To avoid ambiguity, I would distinctly notify to the reader that I shall use terms of relative position, as back
and front, before and behind, in describing the organs of the genital cavity, as if they were independent organisms,
and not parts of the entire animal. Thus, if I speak of the uncus, 1 might describe it as narrowing behind the tip ;
of the smaller seaphium-tooth as behind the larger; of the aristae of the cheeks as directed backwards; of the sheath
of the penis as running back to its base. Now, in each of these cases, the direction intended is certainly from the
tail towards the head of the insect; yet, limiting our attention to these orgaus.it would lie most unnatural and
misleading, to speak, for instance, of the tip of the penis as its hinder portion, of the basal bulb as its front. In
speaking of the abdomen generally, or of its segments, I shall use the ordinary terms, as above.

SECOND SERIES. — ZOOLOGY, VOL. II. 39

274 MR. P. H. GOSSE ON THE CLASPING-OliGANS

a triangle, of which the base is the edge of the seventh segment, the sides are often
incurved, and the point is, in general, prolonged into a strong, horny spine ; e. g., Orn.
Amphrysus, P. Merope, Erechtheus. This point, however, varies much, in different
species, in length, breadth, sharpness, curvature, and direction ; and, as it is one of the
elements of the prehensile apparatus which I am proposing to describe, I distinguish
it by the term uncus, a drag-hook. The term "tegumen," adopted by Dr. White, seems
hardly appropriate for this spinous, often wiry, point, which is certainly in no sense a
cover. But it does express the function of the wide annular part of the segment, from
which the uncus projects ; and in this sense I use it.

Viewed from the side, the tegumen and uncus together often present a very curious
resemblance to the skull of a bird without the lower mandible, e. g. Erechtheus,
Demoleus; which latter, however, is simulated occasionally by the scaphiwm, so as to
complete the likeness — Deiphontes, Homerus, Childrcnce, and very curiously in Tarn.
Apollo. I occasionally avail myself of this and like accidental resemblances, for more
definite and graphic description.

The extreme point is generally somewhat swollen and flattened, spoon-wise — and this
even when the organ is of a wiry slenderness, as in Bhodifer. Sometimes it is bifid —
Agamemnon, Poly dorus; more rarely even trifid — Argesilaus, Archesilaus; sometimes
it is curved upward — Agamemnon; sometimes vertically downward — Sector, Lycidas,
Agavus ; sometimes it is short, thick, and uncouth — Antenor, Zalmoxis, Bromius ; almost
straight — O. Arruana, P. Hesperus, Bhodlfer; often nearly a half-hoop of very slender
wire, displaying a strength, elasticity, and polish, almost of steel — Vertumnus, Macedon,
Turn as. The receding margins are often turned up to a height that considerably exceeds
the plane of the dorsal surface, which then becomes transversely concave — Mayo,
Machaon, Zalmoxis, Merope ; this surface is sometimes indented Avith a trilateral
(Antenor) or quadrilateral (Agesila/is) depression. Often it is clothed with a median
ridge of long and close-set hairs, which project horizontally, and, so to speak, thatch the
cavity and its organs — e. g. Agamemnon, Codrus, Homerus, and, signally, Diphilus, and
the abnormal Hector group, though here there is no uncus. More commonly the stiff
ridge-hairs are evect--Erechthem, Bhetenor, Merope; or, at first erect, and then arching
forward — Menestheus, Machaon .

When the valves arc perfectly developed, as in Ornithoptera, the uncus is projected
immediately under the line produced by the meeting of the dorsal edges of the valves ;
and then the median ridge of long hair helps to fill the blank space between.

The inferior surface of the uncus usually bears two conspicuous laminte descending
vertically from it, more or less deep, which take their origin insensibly near the point,
and increase in depth gradually — Erechtheus, Bhetenor, Helenus, Merope, following the
curvature of the margius till they are lost in the descending rami ; but in some cases this
keel begins at some distance from the point, and with its full depth — Deiphontes, Homerus.

On each side, near the spot at which the rami of the keel can no longer be distin-
guished, a conspicuous horny tooth often projects horizontally from the edge of the
segment — Helenus, Tv/rnus.

The uncus is apparently composed of nearly pure chitine, whose varying density is

IX CEETAIN LEPIDOPTERA. 275

perhaps indicated by its colour, which may be pale fawn, or rufous, or deep sienna-
brown, or even almost black, always with a polisbed surface.

4. The Scaphium.
By this term I indicate a curious organ, which I find almost always present, of com-
plicate structure, apparently having an intimate relation to the uncus, and even, in
general, organically united with it, but yet occupying its proper place even in those rare
cases in which the uncus itself is obsolete, as in Orn. Brookeana, P. Podalirius, and
Bathycles. Indeed in this last-named it is seen to great advantage ; and still more so in
Mayo, the beautiful representative of Poly inn est or in the Andaman Isles ; and most of
all in the African Merope.

The instant we have removed either of the valves from one of these Butterflies, our
attention is arrested by a great mass of shining white tissue, occupying the chief place in
the genital cavity, and projecting far into its area. We trace it up to an intimate and
apparently organic union with the lower surface of the uncus near its origin, far back in
the palate of the bird's skull (if I may use my own comparison on p. 274, supra),
whence it descends and dilates, sending large angular lobes back even into the abdomen
(Bathycles), but chiefly developed forward, like the contiguous organs. I have likened
this prominent part of the organ to a lower jaw (rather mammalian than avian , however)
— Machaon, Turtms, Erithonius, Bemoleus, Maceduu ; I have likened it also to a boat —
Mayo, Pammon.

Of the function of this conspicuous organ I cannot speak with certainty. I adopt a
distinctive appellation for it (presuming it to be important and undescribed), which
leaves function untouched, and looks only at the accidental resemblance alluded to —
<TA.a*fn), (jKaty'iov, a boat.

Where it is most perfectly developed, e. g. Mayo, the sides swell out like the bows of a
ship, while the mesial portion is abruptly thinned away to a deep projection, like a cut-
water and a keel. The upper surface forms two dilated margins (the gunwales of the
boat), with a deep sulcus between, in which in some cases — Vertumnus, Erechthens,
Bheleuor — the uncus lies. More usually there is considerable vertical space between
the uncus and the scaphium-margins. These very generally bear (here the simile shifts
back from the gunwales of a boat to a lower jaw) a compound armature, most difficult
to explain, but in which surely lies hid the key to the explanation. In some clear
examples it mimics the double molar tooth of a mammalian jaw, the outer usually much
more distinctly developed than the inner, often rising to a strong, conical, produced
spine, which may take the form of a straight blunt pin or peg — Memnon, Nireus ; or that
of a canine tooth, erect, acute, recurved — Mayo, Pammon, Arclurus, Macedon, Thoas ;
horizontal, recurved — Machaon, Ulysses ; horizontal, decurved — Bhodifer. Both may
appear as two equal, stout, polished cones — Momerus ; more commonly the secondary is
reduced to a mere conic knob, or is even obsolescent ; occasionally a third supernumeraLy
tooth or knob appears — Mayo, Rome r us ; and not infrequently neither can be detected,
e. g. O. Arruana.

Besides these teeth, there is another kind of arming : the gunwale-like margin rises

39*

276 ME. P. H. GOSSE ON THE CLASPIXG-ORGAXS

into an elevated ridge, which is thinned off to an edge, and is then cut into a number of
parallel erect membranous points or bristles. When the double teeth are wanting this
aristate crest appears to supply their place, as in several of the Omithoptera ; but yet
occasionally both are seen, and then the arista? are more baseward — Hrechtheus.

But the most remarkable form of the scaphium known to me is that of Merope, in
which the teeth are stout, broad, and most elaborately notched and bristled on their
edges ; for the details of which, with other points of interest, I must refer to my account
and figures of the species, ut infra.

The lower parts of this organ are obscure, descending and receding towards the abdo-
men. Sometimes it is patent that the descending rami do certainly embrace the basal
region of the penis, and appear to unite again below it, but are not, I think, organically
united with it — e.g. O. Amphrysus, P. Rhetenor, Bathyoles, Agamemnon. But such
connexion of the scaphium with the penis cannot by any means always be affirmed.

The usual appearance of the organ is opaque white, smooth, shining, like polished
ivory, often very pure, particularly the keel. Sometimes, however, it becomes partially
or even wholly pale, or even dark, hrown — Podalirius, Zalmoxis, Lycidas, JIacedon —
as if the chitinous element pervaded it, which microscopic examination confirm-- .

By the same test the substance appears to be mainly muscular — a conclusion to which
I had already come from the armature of the double teeth. It was impossible to look
upon those formidable weapons, in Merope, for example, without inferring that the
compact mass upon which they are seated must be muscular, or they would be useless.
But I suspect that in this organ reside the muscles which move the uncus — perhaps also,
at least in part, the valves and the harpes, for which offices they would need to be
vigorous and massive.

Many modifications of the form and conditions of the organ occur in different species,
which will be described seriatim f.

On the nature, or even the existence, of the scaphium, I have little help from my
predecessors. Herold, in Pieris BrassicceX, does not distinguish it from the uncus,
which I find quite distinct, both being present, but contiguous. He, however, con-
founds the two organs under the name of the triangle-piece (" Triangelstuck "), which
he cites from De Geer. The minute projection from beneath the point, which he takes
to he the end of the rectum, is surely the keel of a small scaphium. The anus is, as I
believe, in the middle of that transverse line at which the muscular base of the scaphium
is united to the palate-like under side of the uncus §. (See P. Arcturus.)

* The scaphium of P. Machaon, macerated and torn from its uncus, showed, at the upper part (the gunwale), a clear
yellow-brown tissue, with smooth edge, and minute bristles growing from the surface. This ran off into colourless
transparent tissue, decidedly fibrous, several of the bundles being formed of longitudinal fibres, which were more or
less covered with thin lamince filled with excessively minute but perfectly distinct regular parallel oblique strur.

t For some further researches into the structure and relations of the scaphium, made after this memoir was
handed in, I beg to refer to the account of Omiihoptera Remus, infra.

X Entwickelung der Schmett. xiii. tab. iv. figs. 3-6.

§ The fact that the uncus is tubular, as may be clearly seen when it is forcibly broken across, suggested the
thought that perhaps itself might be the termination of the intestine. But very exact and repeated examinations

IX CERTAIN LEPIDOPTEEA. 277

Neither Burmeister nor Siebold makes any allusion to the scaphium. De Haan's
slight mention of " lateral plates " has been already cited. From his figure, I suppose
he alludes to the scaphium, by this phrase. In the numerous figures of Dr. White's
Memoir *, minutely beautiful, and carefully executed, as these are, I find no trace of
the organ; and I think that probably it is peculiar, or nearly peculiar, to the true
Equites. I say " nearly peculiar," because, though Dr. White finds it not, as a separate
organ, in the European Pieridae, yet I hope presently to show that unequivocal traces
of it are to be found in certain members of the family.

5. The Penis.

This organ forms, strictly, no part of my subject, which is not the function of gene-
ration, nor the organs that perform it, but certain prehensile apparatus that are ancillary
to the performance. Yet, as this member is so essentially the centre around which the
whole apparatus waits and serves, and as it forms so conspicuous an object in the
accompanying designs, I can scarcely avoid giving some account, at least, of its varying
form and position.

The penis is usually seen, when the valves are opened, in the lower half of the genital
cavity, in the form of a column, more or less cylindrical, varying much in diameter and
in length, projecting from the interior of the abdomen. In some instances it is so short
as not to appear beyond the walls of the eighth segment, e. g. Vertumnus, Lycidas; but
it is doubtless capable of being protruded, to a certain extent, by its proper muscles f . In
others it is so long that it cannot be contained within the cavity, but projects beyond
the margins of the valves when these are shut. In Merope, JJealecjon, Hesperus, Iihe-
teuor, the tip just reaches the edges of the valves, or exceeds them in a slight degree;
in Podalirias, Bromius, Macedon, and still more in Bathycles, the protrusion is remark-
able; but, in the Coon group, including the beautiful Rhodifer lately sent from Anda-
man, the penis projects from the hinder extremity to an extraordinary degree, like a fine
steel wire, that quite arrests the attention as we see the insect in a cabinet. No less
notable is the extreme slenderness, and, at the same time, the strength and elasticity, of
the organ in the just-mentioned species.

As to position, the penis appears so high as almost to be in contact with the scaphium
— Tamils ; again, almost on the very floor of the cavity — Deipltontes ; and, in some
species or other, in every intermediate stage.

have convinced me that the extremity of this organ is solidly ehitinous, and absolutely imperforate, at least in
0. Remus and 0. Haliphron. But, when the uncus is removed, there is clearly discerned, under the very middle
of the tegumen, a circular orifice leading from the abdomen into a groove, which is hollowed along the median line
of the scaphium-sulcus. In dried specimens the intestinal convolutions are reduced to an undistinguishable mass of
yellow dust. Yet, on one occasion, I detected (or fancied) a thread connecting the mass with the orifice just described.
It had, however, vanished in an instant.

* Op. tit.

+ In an abdomen of 0. Haliphron I carefully laid bare the penis till I traced the base of the chitmous cylinder
originating in firm muscular tissue, in the ventral arch of the seventh (that is, the ante-penultimate) segment. But
in P. Merope, I laid bare the cylinder to its base in muscular tissue, in the sixth, or even the fifth segment.

278 ME. P. H. GOSSE ON THE CLASPING-OTJGANS

A curious variation lies in this circumstance, that, while the proper condition is that the
darker, more chitiuous side of the cylinder, that which terminates in the finger-point, is
uppermost, there are numerous examples in which this order is reversed, and that side
is undermost, as in several of the Omithopterce, P. Agamemnon, Agavus, and others.
More rarely this (normally) upper side is only partly turned round ; so that the odd
notion is suggested, that the penis has the power of turning round on its long axis at
the will of the insect.

It is also variously subject to curvature, often simply arched, with the tip downward,
by the lengthening of the middle line dorsally — Mayo, Rhodifer, Polycaon; sometimes
with the tip upward — Pammon, Hesperus; sometimes to either side — Memnon, Ver-
tumnus, Agesilaus; sometimes it is thrown into a double (sigmoid) curve — Rhetenor.

The form is ordinarily cylindrical and tubular — Pammon, Zalmoxis; sometimes
spindle-shaped — Podalirius, Agamemnon ; the upper (denser) line very frequently pro-
duced into a finger-like point — 31ayo, Nireus, Macedon, of varying length. The extre-
mities of the sides usually expand, so as to make a wide, very oblique trumpet-mouth —
Helenus, Erithonius, Ulysses; and sometimes project beyond the general level into distinct
lips, or a sort of prepuce, often corrugated — Erechtheus, Mayo. In P. Semperi the
expanded orifice is of extraordinary dimensions.

In its structure the penis seems to be formed of two distinct tissues, often wrell defined.
Tbat side of the length which is normally the upper is a cylinder of more or less dense
and polished chitine — Hector, Podalirius, which is lined by another tissue of soft, pulpy,
white matter, of shining surface. The former is often (not always — P. Sesostris) divided
along its underside throughout; and the white tissue then exceeds its bounds — Hector,
Macedon, Podalirius ; sometimes notably, Codrus. This white element has other very
singular developments, for particulars of which I must refer to my detailed descrip-
tions and figures, particularly the curious globe protruded from the wiry organ of
P. Doubled <t if i, and the still more curious and thought-evoking cases of Orn. Rlchmondia,
Haliphron, and Darsius*.

The variety of position and direction assumed by this organ, as has just been described,
seems to imply almost unlimited freedom of motion upon its basis. Yet, now and then,
we discern surrounding and, as I presume, supporting and restraining ligatures. That
the seaphium seems normally to send, inward and downward, fleshy branches, which, in
certain cases — Bathycles, Mayo, Agamemnon (all which see for particulars), surround th€
penis, I have already mentioned. In P. Erechtheus a thick white ligament rises from
the floor of the genital cavity, or from each of the chitiuous knobs which afford interior
hases to the harpes, and, thrown over the penis, not far from its tip, appear to bind it
downward. But I have seen nothing like this elsewhere.

In P. Doubledayi and Rhodifer (and also, on De Haan's authority, in Coon) the lonf

* In the great Asiatic Moth, Atiaeus Atlas, where the organ has exactly the same form and position, I find, <lurin>-
life, that the terminal tissue, which fills the expansion, is very definitely separated from the brown chitiuous waUlj
of the tube ; it is pellucid-white, soft, and pulpy.

IN CERTAIN LEPIDOPTEEA. 279

wiry penis issues from a conical sheath, which closely envelops it, as if an organic part
of it. This does not appear to he a process from the scaphium.

Finally there is that remarkable dividing wall, of hard, horny chitine, which I describe
and figure in Erithonius, perforated in the middle, manifestly for the egress of the penis,
yet which, strange to say, I have met with in no other species *.

Tue Purpose of the Structural Variations.

If it be asked, What i§ the definite purpose, in the economy of the creature, of this
extreme variation ? I am obliged to answer, I do not know. That, viewed in the large,
the object of all these organs that crowd around the termination of the male abdomen is
the firm grasp and sustained retention of the female abdomen, in the delicate and most
essentially important function of reproduction, is sufficiently evident. But why the
diversity of detail ? Why would not one good and adequate form suffice, again, and
again, and again, subject to no more variation than are the antenna?, for example, or
the tarsi ?

It naturally occurred to me, very early in these researches, that every peculiarity in
the prehensile organs of the male would have a correspondent peculiarity in that
part of the female body which they were formed to grasp ; and I eagerly turned to the
examination of the female abdomen. But the repeated search left-, and still leaves, my
Question — cui bono ? — without an answer.

The position of these organs, and their relation to the surrounding parts, when in situ,
show indubitably that it is the exterior of the final segments of the female abdomen,
that are seized in coitu. The harpes and the uncus are certainly not intruded into the
female abdominal cavity. Thus the field of search is at once limited to the outer surface,
nrom the very nature of things ; and this is confirmed by the occasional clogging of the
|iarpes with body-scales, as I have already noticed. It was then sufficient to denude the
female abdomen of its clothing-scales, without disturbing the parts.

But though I did this with species after species, selecting those whose males have
Inversely-armed harpes, no solution of the inquiry was obtained. The females of Pa pit "to
\Thoas, Poli/damas, Anchisiades, Ileteuits, are, as to the point in question, mutually

I * The genital organs unci their accessories have been minutely described and exquisitely figured in another Order of
cts. The question maybe asked,— What homology (or analogy ?) exists between the organs herein described

me and those of Trichoptera described by Mr. McLaehlan? I have examined every figure in his great work, and
■onfess that I know not how to institute any satisfactory comparison with those parts iu Papilio. It is just possible
hat the " dorsal process," in such forms as Ehyacophila, may be equivalent to my "uncus;" but of "scaphium"
L see not a trace. Possibly, too, the " inferior appendages," so largely developed in the same genus, may represent
:he " valves ;" but the resemblance is most evanescent. Tho penis seems formed on a plan wholly different.

In one point my own observations agree with those of the learned author — the remarkable fact that no two species
feem to have the same forms of armature.

My friend has been so kind as to send me his paper " On the Sexual Apparatus of the Male Acentropus " (Trans.
Ent. Soc. 1872). Looking at his figures, I should have been inclined to say, if I had not been told, that they repre-
sented the parts of some Papilio or Pieris. We seem to have the valve (a), with its dorsal moiety, indeed, absorbed
the tegumen projected into an uncus (6 and c) ; and the scaphium (d, e) small, but normal in form, connected basally
with the uncus, and bearing its usual lateral elevations, duly crowned with teeth or the characteristic aristaa. The
parallel is most curious. I should expect some form of harpc lurking within those appt ndices inferiors.

nsi

280 ME, P. H. GOSSE ON THE CLASPING-OEGANS

undistinguishable. Under the microscope, the whole surface, after denudation, appears
pitted with very minute shallow depressions, each with a knob at the bottom. They
are arranged in transverse lines; and those of each line alternate with those of the lines
preceding and following it, but this by no means witli mathematical precision. These
pits are avcII known as the receptacles of the clothing scales. Possibly, indeed, they may
have a second use, even to afford hold for the saw-points of the male harpes ; and the
clogging scales, carried away by the latter, had been, no doubt, displaced by the saw-
teeth in securing their grip-hold.

At the posterior margins of the seventh and eighth segments, these surface-pits appear
closer together and larger than in other parts of the female abdomen.

The surface of the ninth segment, which consists of the valvules of the vulva, is
peculiar — not clothed with scales, but cloth-like, being composed of hairs nearly erect,
placed exceedingly close together, and, though very fine, so short as to be, in places, of
a length not more than five or six times their thickness, as may be discerned when
under the microscope we view the very edge of the rounded surface against the light.

Now, though it is not improbable that these last-named parts, the exteriors of the
valvules of the vulva, may be the spots ordinarily grasped by the apj)roaching harpe-
tips, yet I can discern nothing iu the nature of their surface, nor in that of the proximate
surfaces, to distinguish one species from another — each surface, in one, exactly correl
sponding to the same surface in each of the other species, whose females 1 examined.
And so there is nothing, so far as I yet know, to account for the astounding variety in
the harpes of the males.

Yet, if I see a number of keys, of very minute and elaborate workmanship, all different,!
I cannot doubt that every one is intended to fit some special lock, though I have not
examined the wards ; and this conviction is the stronger, the more varied, and the more
complex are the keys. We cannot withhold a hearty assent to the conclusion of one of
the most eminent of modern physiologists, who, speaking of these organs in the class of
insects generally, says, "They prevent allied species from producing bastards by adulterous
connexions ; for the hard parts of the male correspond so exactly with those of the
female, that the organs of one species cannot fit those of another"*. And Leon Dufour
speaks of them as " a guarantee of the conservation of types, and a safeguard for the
legitimacy of species" f . But I should like to see these axioms demonstrated.

What anomalies remain to be discovered, I cannot guess by the closest scrutiny of
the lovely wings spread out in our cabinets. The strangest deviation from normal form
that has occurred to my notice, is the unvalved abdominal apparatus in the beautiful
Papilio Rector of India. It is most aberrant ; but what hint of this is conveyed by the
gay body and wings ? Who could conjecture, by looking over a cabinet, that Ascalaphm
and Helen/is are so alike in their genital armature, while Hector and Agavus are so
unlike ?

On the Harpe-tvpes.

As the harpe appears to be the leading organ in the prehensile apparatus, the ruosll
fully elaborated and the most varied, I attempt a grouping of the different forms of this

* Siebold, Comp. Anat. § 354, note 2. f Ann. des Sci. Nat. i. (1844), p. 036.

IN CERTAIN LEPIDOPTERA.

281

organ. ^ The result is an order of succession which, it must be confessed, sets at nought
that which is founded on the form, neuration, and colours of the wings. It is not proposed
as, in any sense, a natural arrangement of the species, but as a help to reference and
memory, and as a means of comparison of the varying conditions of this special
organ. The groups themselves are loosely defined, and run into each other.

It must be recollected that the armature of only sixty-nine species, out of the four
hundred and more described PapiUones, is here recorded. A fuller examination will
doubtless bridge-over many gaps, and supply other forms.

Orntthoptera.

1. A broad arm ending in a bent spiniferous hand.

Arruana. Haliphron.

Pronomus. Darsius.

Ricliniondia. Rhadaniantlius.

2. A long wire, springing from a square plate.

Amplirysus.

3. A small hollow shell, with toothed edjyes.

Remus.

Heliacon.

Heliaconoides.
Brookeana.

Nephelus.
Cliaon.
Mencstheus .
Pammon.

Agavus.

Papilio.

4. A shaft with securiform head, and generally a tooth beneath.

Memncm. Schmeltzi.

MaJr°- Polyrnnestor.

Erechtheus. Proteuor.

Rhetenor. Ascalapkus.

Deiphoutes. Helenus.

5. A sword with a saw-edge, or pointed, and set edgewise.

Macliaon. Merope.

Arctums. Brutus.

6. A long two-pointed blade, not serrate.

Hornerus.

7. A long pointed style, not serrate.

UIysses. Phorcas.

8. A broad thin plate, with thickened edges, especiaUy the ventral ; variously toothed.

Tnoas- Bathycles.

Turuus.

Polycaon.

Axion.

Hesperus.

Dcnioleus.

Agamemnon.
Erithonius.
Anchisiades.
Vertumuus.

Sesostris.

Children^.

Lysander.

Euryleon.

Ergetcles.

9. Tlie plate obliquely truncate, serrate, or running into broad points ; with deviations
Macedon. Latreiliei. Doubledayi.

?r°™- Semperi. Rhodifer.

Podahnus. Lycidas.

SECOND SERIES.— ZOOLOGY, VOL. II. 40

282 ME. P. H. GOSSE ON THE CLASPING-OKGANS

10. The points produced, into long styles.

Archesilaus. Zalmoxis.

11. The plate complicate, sending off many laminaa and points, often serrate.

Policones. Agesilaus. Codras.

Ucalegon. Parmatus.

12. Rotundo-triangular, with a spoon-like knob, usually studded with points.

Nireus. Polydorus. Hector.

Dipliilus. Antenor.

Specific descriptions.

Ornithoptera arruana, Feld. (Plate XXVI. figs. 1-3.)

In this grand butterfly the valve is of unusual size, measuring "3 inch in length, and
"33 in breadth. Externally it is uniformly brilliant yellow ; internally, rich velvety
brown. Its outline is rotundo-triangular, with the dorsal side cut off obliquely at the hasal
moiety, and the ventral side correspondingly produced. The ventral margin is broad
and flat, or rather slightly channelled ; there is a projecting blunt tooth, about one third
down this margin ; and the whole is surrounded by a narrow fringe of hairs, blackish
and dull on the dorsal, brassy and glittering on the ventral edge. The ventral margin
forms a wide flat shelf throughout, before the level descends to the concavity.

At the very base of the valve begins an ample Iwrpe, in the form of a broad plate of
glittering chitine, narrowed slightly in the upper part, and then expanding into a spoon-
like disk, of which the terminal edge is flat, oblique, and beset with minute black
glittering, curved, acute spines, the points directed ventrally, arranged, inexactly, in
transverse rows, of which there are about seven on the dorsal edge, diminishing to one
before the ventral angle is quite reached *.

The hurpe consists of two surfaces, as if a tube had been flattened ; and, besides this,
the upper surface is hollowed both immediately below the spincd disk, and in the basal
moiety. The terminal portion, about one fourth of the whole, is disconnected with the
lining membrane of the valve, to which the remaining portion adheres ; but the whole
can be lifted, and separated from the valve by delicate manipulation, and placed on a
glass slide for higher amplification. The component chitine is now seen to be quite
transparent, of a rich yellow-brown colour, deeper at the edges, and specially at the
extremity, where the spines are glittering deep black. All over the body of the harpe
ramify many irregularly sinuous lines, like rivers in a map ; and these, I found, by an
accidental fracture of the structure, are not on the outside, but on the inside of the walls.
When the valves are closed, the pah' of spinous disks meet accurately at the very tip of
the uncus, as may be seem with the left valve in PI. XXVI. fig. 3.

The uncus is a wire of extreme slenderncss, black, polished, very slightly curved,
ending in a point, not sensibly dilated, not very acute ; the basal portion of the upper
surface bears a ridge of close, stiff, black hairs, which stand nearly upright.

* It must be remembered that, as in many otber eases, this is the careful description of an individual ; the
minuter features may not be repeated in every example.

IN CERTAIN LEPLDOPTERA. 283

The scaphium is very difficult to understand without dissection, to which I have been
reluctant to resort with so valuable a species, contenting myself with careful study of
the parts in situ. It seems to me, after much consideration, that the scaphium proper
is reduced to a slender " boat " running off to a point under the tip of the uncus to which
it seems attached, above the triangular basal expansion, by fibrous processes on each
side. I can see nothing answering to the ordinary " double teeth." The keel, on the
other hand, seems to be unusually developed into the large bagging body which hangs
below *.

In the lower part of the cavity is seen the penis — here a moderately thick column of
polished brown chitine which projects almost horizontally from the abdomen to the
edge of the valve, and terminates in a long drawn point, just before which it sends
off two expanded foliations on the upper side (the organ being reversed), which are of
thinner substance than the tube, and which face each other.

Towards the base of the genital cavity, projecting from one of the lower dilatations of
the scaphium, I find, in the forms of this species, particularly distinct in 0. Richmondia,
two short sharp polished spines on each side of the penis, so placed that the four are in
a line with it. I cannot suggest the use of them ; their position seems to preclude
prehension.

Obnithopteba pbonomus, G. R. Gray. (Plate XXVI. fig. 4.)

The harpe is not distinguishable from that of 0. Arrucma, except that the armed disk
is a little wider. In removing the valve, the harpe detached itself, together with a
portion of the lining-membrane, so that I could lay it, quite alone, on a glass slide
for the stage of the microscope. The black shining teeth are about fifty in number,
arranged exactly as in Arrucma, and interspersed with stout bristles or straight spines,
all projecting at nearly a right angle to the plane of the stem. The teeth are some larger
than others, irregularly crowded, decidedly curved, stout, pointed, black, glittering in the
light; the chitinous substance of the disk and stem translucent, gall-yellow by the
transmitted light, varying in depth of tint, according to the thickness, and, as I think,
according to the density, of the material.

The uncus differs much from that of both the preceding and succeeding forms. It is
rather short, not uniformly curved, but bent down near its middle with a kind of knee ;
it widens rather rapidly to the base ; and the vertical rami, which I will call the keel of
the uncus, increase in depth rapidly. The uncus is not why.

I could find not the least trace of a scaphium, in the single specimen that I examined.

The penis, reversed, was abruptly inclined from near the base downward ; so that
it lay in the very angle between the valves, and was so long that the tip was fully
seen without, pointing obliquely towards the head of the insect. The expanse of the
lips was empty to a considerable depth ; the median line channelled beneath, as if the
integument there were thin, and had contracted in drying.

* Since the above was written I have made examination of the specimen in another mode. Having softened the
parts by moisture, I stretched them apart with needles, and thus satisfied myself that my explanation hazarded
above is the true one.

10*

284 ME. P. H. GOSSE ON THE CLASPING-OKGANS

The scapkium may have been accidentally lost ; but tbe peculiarities in the uncus and
penis are considerable, and, if these prove to be constant, are good distinctions between
this form and its fellows.

Ornpthoptera richmondia, G. R. Gray. (Plate XXVI. fig. 5.)

Here the harpe diverges from the preceding forms. The disk is considerably wider
and shorter, without any sensible diminution into a neck, and it shows a distinct
tendency to that form, which I have compared to an open hand bent back from the wrist,
conspicuous in the black and yellow species of the genus, but which is scarcely seen in
the Priarnoid races. The teeth are crowded along the truncate extremity (the " fingers"),
and within (on the " palm "), and more on one side than on the other ; but, whereas, in
.Pronomus and Arrucma, the main crowding is at the left corner (as looked at), here, in
Richmondia, it is at the right. The teeth, though minute, are seen, when viewed at the
proper angle, to be curved, acute, glittering spines ; their number is about fifty.

The uncus and scaphium are as in Arrucma ; but the former is somewhat more curved.
The penis also agrees fairly ; but the expansion (in one examined) was much less
corrugated, and had a broader, more truncate point. The organ was manifestly empty,
as my lens reached for some distance up the interior. But another example was
full, to swelling, throughout its whole length, on the inferior side, with the singular
white pulp, which also had filled the expanded lips, and reached to some distance
beyond. Moreover, as I had removed the valve from this example, with a slight jerk,
there had been projected from it a compact sinning white globule of this same matter, as
large as a rape-seed, which I have every reason to believe had originally borne the same
relation to the penis as the like knob represented in my illustration of P. Doubledayi in
PI. XXXII. fig. 18.

*■&■

Ornithoptera Halipiiron, Boisd. (Plate XXVI. figs. 0 & 7.)

Outline of valve rotundo-triangular, the sides much rounded, occasionally so as to
approach the circular form. Prom the tip projects a minute horny point, which bends
inward, towards the opposite valve. Interiorly, the margins slightly overlap, and the
edges are sparsely set with short hairs. The concavity is smooth, not polished.

The harpe occupies the middle of the valve, and is a flattened tube (?) of rufous, trans-
parent, highly polished chitiue. It springs from a wide base, like the trunk of a tree,
which, on one side, adheres to the base of the valve, and, on the other, to a stout angular
projection at the bottom of the abdominal cavity. Thence it pursues the concavity of the
valve, narrowing as it proceeds, adhering to the lining-membrane, and scarcely elevated
above it, till about two thirds of the length of the valve are reached; here the chitinous
column is detached from the lining, and rises sensibly from it. It is here narrowed to
a sort of neck, then abruptly widens to a broad ovate disk, studded all round its margin
with stout sharp glittering teeth, fifteen or sixteen in number, all of which point inward
Avhen the valve is closed.

The armed disk is abruptly bent backward, and a little to one side. It may not
inaptly be compared to an open hand, bent back upon the wrist of a naked arm, as far

IN CEETAIN LEPIDOPTEEA. 285

as it will go, the upcurved fingers representing the hooked teeth, and supplemented by
a ridge of teeth running obliquely across the wrist itself.
The uncus. — When the two valves are closed, as in life, the hand-like disks approach
udtther; and, between the facing rovrs of teeth, descends the point of the uncus.
This organ is not so wiry as in the preceding species, but rather suggests the blade of a
clasp-knife, projecting, with a wide triangular base, from the dorsal arch of the eighth
abdominal segment, having a moderately curved upper edge, and ending in a sharp
hooked point. What looks like the blade is truly the side of each of two vertical rami,
which, supporting and strengthening the structure, pass off laterally into the walls of the
segment. The whole is formed of bard horny chitine, dense in texture, dark in colour,
and glittering in surface, even as are all the prehensile organs which form the subject
of the present memoir. Sometimes the edge of the uncus-blade has a conspicuous notch ;
but the minuter details of all the organs are not absolutely identical in different examples
of the same species. One example had a minute supernumerary point just under the
proper point ; in another, the triangular base, viewed vertically, was distinctly notched
on each side, just where it narrowed to form the blade.

The scaphium is large and plicated, white, shining, firm, the keel always prominent,
but in some more than in others. As in 0. Amphrysus and Rhaclamanthus, the rounded
lobe on each side rises to a thin crest, divided into notches, each of which is crowned with
an erect bristle, rising above the level of the uncus. The horizontal point of this lobe,
somewhat condensed in texture, seems to represent the " double tooth " of other
species. The median groove receives the blade of the uncus, as into a haft.

The penis, as usual in the genus, is large, and funnel-like at its extremity ; in position
it is reversed, the more chitinous side, which terminates in a finger-like point, being
downward. In one specimen, this organ displayed a remarkable phenomenon (which,
under modifications, has recurred to me since), which I cannot satisfactorily explain.
Even the unaided eye could discern a thick white mass, connecting the trumpet-mouth of
the penis with the inferior surface of the scaphium. I immediately introduced a drop
of water into the genital cavity, and allowed it to stand awhile. I then removed, with
absorbent paper, the superfluous moisture, and applied a considerable magnifying power.
The appearance now presented, I have depicted with great care, in PL XXVI. fig. 7.
A cylindrical column of soft, subgelatinous, mirk-white substance, of a diameter slightly
less than that of the expanse of the penis, had, apparently, risen vertically from the
interior of this organ, till it had come into contact with the scaphium, against which it
had flattened by the continued vis a (ergo. The lower part of this column was tinged
with the yellow-brown hue of the organ itself. The chitinous lips of the expansion
seemed to have been forced wider apart — split, in short — by the protrusion ; and, what
seems to me very suggestive, certain lines of very pale chitine-colour, agreeing in form
with the outlines of the chitinous lips, looked as if the lining membrane consisted of
several coats, and as if these had been protruded, successively, one after another, as the
white matter had been pushed out !

But, what is the nature, what the function, of this white jelly-like matter ? Is it an
integral part of the organ ? Is it a secretion from it ? Is it the semen, under novel

286 ME. P. H. GOSSE ON THE CLASPING-OEGANS

conditions ? Its firm consistency, even when softened by the absorption of water, and
its condition, when dry, of a firm, solid, shining white mass, seem inconsistent with such
a conclusion. It may, possibly, not be without relevance, that the spinous disks of the
harpes were, in this specimen, unusually clogged with meconium and scales, suggestive
of a recent coitus at the period of capture.

But I would rather incline to seek analogy with cases, not very infrequent, in which
the penis is manifestly separable into two longitudinal portions— a split tube of brown
chitine, and a pulpy white mass occupying, and more than filling, it *. (See my
figures of Hector, Codrus. Agamemnon, Erithonius, Podalirius.)

Ornithopteka Darsius, Gray. (Plate XXVI. figs. 8-11.)

Valve almost exactly as in O. Haliphron, save that the terminal finger-point is a little

longer, and spathulate. The interior has a rather wide space within the ventral margin,

level with it, before it abruptly drops to the cavity.

The harpe also has a general resemblance, but is narrower, the disk more truncate,
more excavate, furnished with eight or ten stout and sharp teeth, all marginal. Viewed

sidewise (PL XXVI. fig. 10) it takes something of the form of the human arm and open

hand, but less bent back upon the wrist.
Uncus, much as in Haliphron, viewed laterally ; vertically, it is more slender, with a

slight dilatation just behind the point.
Scaphittm, ample, with a deep thin keel, cjuite white.

The penis, in the individual examined, again presented the very remarkable pheno-
menon which I have described in Haliphron. The organ was reversed, slender in the
column, with a very ventricose, one-sided trumpet-mouth, that had apparently been split
open by the extrusion of a great globose mass of the white substance. Here, as there,
several successive laruinse had, apparently, been thrust out, in various degrees, just as I
have there represented, all showing definite angular points. The white substance, so
much resembling the albumen of an egg coagulated, was, when I first looked at it, in
contact with the scapiiiuin. It absorbed water : after which I could easily, with a needle,
detach minute portions for microscopic examination. These, diffused in a drop of water
on a slip of glass, showed no trace of organization, but resolved themselves into irregular
atoms of amorphous matter. Fig. 11, PL XXVI., represents this penis viewed nearly
from the front, when most of the white matter had been removed.

This phenomenon, which has occurred in three distinct species of Ornithoptera, viz.
0. llicliiuoiidia, Haliphron and Darsius, appears to me very curious and puzzling. I have
assumed that the strange substance is extruded ; but I am not sure. The splitting of
the walls of the penis, or, at least, the forcing apart of what, if expansible and separable
before, were normally closed, certainly suggests a thrusting-in from without, rather than
a thrusting-out from within. Yet this would be, a priori, most unlikely. Whence
should come the matter ? Why should it be pushed into this organ, when it does not
pervade and fill the surrounding parts ? This thought must be dismissed ; and the

* Dr. Burmeister (Mun. Ent. 1S3U, p. 329) speaks of the turgescence of the organs of generation, before and during

copulation.

IN CERTAIN LEPIDOPTEEA. 287

appearance of the successive coats pushed out, one beyond the other, as well as the
flattening and expanding of the white column against the bottom of the scaphium, in the
Haliphron case, seems to me overwhelming evidence that the movement has been from
within.

If we compare the penis as it appears in Rhadamanthus, Amphrysus, and Armaria
(see the figures on Plates XXVI. and XXVII.) with the same organ in these three cases,
or, better still, the penis in one specimen of Haliphron with the same in another, the one
filled, and burst, with this white matter, the other normal in form, brown in hue, empty
at the lips, without the slightest trace of white matter, the thought occurs that possibly
we find here not specific differences, but only two interchangeable conditions. The
shining, expansile, open and empty lips may give the ordinary passive condition ; the
white pulp pervading the tube (see Hector, Plate XXXIII. fig. 31), filling the orifice,
and expanding it to overflowing, and even to a breaking of bounds, may give the con-
dition when the "furor igncns " is raging — may be analogous (I do not say identical,
seeing we have to do with bloodless invertebrata) with erection and emission in the
higher vcrtebrata.

Observations carefully made, in hac re, in individuals taken m coitu, as well the female
as the male, might be of great interest and value.

Supplementary Note to Ornithoptera Darsius.

The globule of white substance from O. Rickmondia, I submitted to my friend Pro-
fessor Gladstone, F.P.S., Pres. Chem. Soc, who has favoured me with the following
report.

" On examining your little particle, I have obtained a more satisfactory result than I
had anticipated. Though the piece of white excretion was no larger than a small pin's
head, I have been able pretty well to determine its constitution. It was quite hard, but
easily pulverizable, and consists mainly of earthy phosphate and some fatty or oily
matter.

" The phosphate melts, when strongly heated, like the ' fusible phosphate ' which is
common in urinary calculi : but the quantity was so minute that I cannot say, with
certainty, whether it contains magnesia, as well as lime ; but my impression, from the
chemical tests applied, is, that it does consist of both.

" There is a considerable proportion of organic matter mixed with this phosphate.
Ether dissolves it out; and, on evaporation, it is obtained again in what, under the
microscope, appears as oily drops.

" No trace of uric acid could be detected."

According to Leon Dufour and other anatomists, the urinary organs in insects " always

consist of tubes which are inserted in the pylorus and terminate caecally The

granular contents of these vessels .... flow gradually into the digestive canal. Thus
excreted they accumulate in the colon, and are evacuated with the ftpces, or separately

as a turbid fluid With the holometabolic Insecta, the urine is evacuated

isolately when they approach the completion of their pupa state. It is well known
that the Lepidoptera, when bursting from their pupae, emit a considerable quantity of

288 ME. P. H. GOSSE ON THE CLASPEVG-ORGANS

urine, of a variable colour." (Siebold's Comp. Anat. — Burnett's Transl. 1854, page 441
and note.)

If tbis statement is correct, tbe penis is never used for tbe evacuation of urine: urine
being identified with that copious fluid which Dr. Burmeister compares to tbe meconium,
and which I have had repeated occasion to mention in these pages. But the frequent
recurrence of this white pulpy matter within the penis and manifestly ejected from it
appears to show its normal, if occasional, connexion with this organ. I cannot help
thinking of the thick white creamy substance, of an overpowering fetor, which I have
seen copiously discharged from the cloaca of Serpents, which, however, is said to be
the urine, "consisting mainly of uric acid" (Grant, Comp. Anat. p. 631). The oil-
globules may possibly look to a changed condition of spermatozoa in residue of semen
partly discharged from the organ, or to fatty degeneration of some tissue.

More recently, my kind friend Mr. Robert M'Lachlan, F.R.S., has favoured me with
some correspondence on the subject. He mentions to me a fact, of which I was not
before aware, that a mass of spermatozoa is sometimes " excluded by other insects in
dying;" and this on the authority of the Rev. Alfred E. Eaton, M.A., whose words, in
a letter to him, my friend thus quotes : — " Oligoneuria rhenana [a May-fly, common on
the Rhine] discharges either spermatozoic paste, instead of spermatozoic liquid, or else
discharges spermatozoic cysts. I have introduced, into plate iii. of my promised work,
a figure of the lobes of the penis with the two threads of paste, or the tubules (as the
case may be) like transparent vermicelli, partly extruded. But nothing has been pub-
lished about this. In other Ephemeridce, the matter ejected is fluid."

Tbe second of these alternatives is, I think, inadmissible. The matter is not contained
in a cyst, if this term implies an enclosing wall, but always as a mass of paste, unde-
fined, save by the cessation of its own substance.

Yet later, I had occasion to examine several examples of the noble Papilio Gigon, from
Celebes. Within the expanded lip of the penis of one of these — protruding, but not
extruded — there was a rounded mass of the dried white pulp, which also was seen filling
the whole cylinder of the abnormally large organ. I easily detached an atom with a
needle, and transferred it, with a drop of water, to the compressorium of a microscope
(Powell's), subjecting it to graduated pressure with a square of thin glass.

The appearance now was as of an infinite multitude of filaments, most unequal and
irregular in thickness and direction, anastomosing (or else crossing) at various intervals,
and enclosing excessively minute areas, having refractive power. At the edge, where
it thinned off to nothing, these areas resolved themselves into minute flattened corpuscles
of no definite form, but always with outlines irregularly sinuated, never angled or
pointed. These corpuscles, closely appressed in the denser parts, made the bright
interspaces, and their edges the darker network. Among them were a few yellowish
molecules, larger and more opaque, irregularly roundish, of various sizes, and also a
good many clear oil-globules, some few of Avhich last were drawn out into slender tails,
of the same substance.

I have taken great pains to be accurate in describing what I saw (under a ^-in. power);,
but I fear it is not very intelligible. I had never before examined spermatozoa that

IN CERTAIN LEPIDOPTERA. 2S9

had been dried ; but the impression was strong on my mind that this was indeed the
nature of what I was looking at. Thus Mr. Eaton's suggestion solves the problem :
these discharged masses consist of paste mainly composed of spermatozoa.

Ornitiiopteka rhadajianthtts, Boisd. (Plate XXVI. figs. 12-16.)

Valve much as in Halipliron. The harpe is separable (clean, and with ease) from the
parchmenty skin which wholly lines the valve ; the latter itself also separable in the
dry state. It is a long narrow plate of shining, transparent, yellow chitine, not quite
so much dilated at the base, with a cavity sunken in the middle of the basal half, having
abrupt irregular edges. The extremity is slightly oblique, pointing dorsally as in
Halipliron, but not in the least dilated, armed with eighteen stout curved teeth, all
placed at the very edge, except two or three which are submarginal. There are also
a few long black bristles, or spines, scattered over the surface.

The teeth, as in the allied species, project from the plane ; hence, when the harpe
is laid on the stage of the microscope (as at PI. XXVI. fig. 14), an inadequate idea is
formed of the power of the armature. When it is tilted, so as to be viewed edgewise,
the teeth, which had looked like low cones, are seen to be strong and very acutely
pointed semi-crescents, bearing a very close resemblance to the spines on the stem
of a rose (see fig. 15, which represents the four teeth on the dorsal margin, viewed
sidewise).

A second example had the form of the valve, and of the harpe, almost exactly the
same ; the teeth seventeen, with almost the very same arrangement, even to the trivial
circumstance of two on the ventral curve being double.

In a third example*, the tip or disk of the harpe was a trifle longer, and decidedly
bent back (like a hand from the wrist) towards the valve-
cavity, the middle part of the arm being bent up to allow '.:^'"~^ls^^^>
this, so that a longitudinal section of the harpe, viewed Harpe of 0. Ehadamanihus,
laterally in the plane of the valve, would assume this form
(see woodcut).

In the intra-abdominal apparatus there is a general agreement with the allied species.
But the uncus is rather shorter and stouter, with a less graceful curve. The scaphmm
differs much, a difference better expressed by the figures than by words (see PI. XXVI.
figs. 12 and 16). There is a similar aristate crest on each side of the uncus, whose points
rise above its level; and from the bottom of the lobe projects a long horizontal
black tooth on each side. The scaphium-keel is unusually developed, projecting with a
sinuous bend, and having a slightly expanded, corrugated point ; its surface bears
many irregular, but symmetrical prominences, hollows, and folds of the firm, shining,
white tissue, which, here and there, deepens into brown or even black — indicative, I
presume, of a more copious supply of the chitinous element to those parts.

" < tftcn my observations have been limited to single examples of each species : wherever, as above. I have had
opportunities of comparing other specimens, I gladly record the results. Their cumulative tendency is to show that
the features of the prehensile organs, and especially of the harpe, are constant, and characteristic of species, not
varying with individuals.

.SECOND SERIES. — ZOOLOGY, VOL. II. 41

290 ME. P. H. GOSSE ON THE CLASPIjNG-OEGANS

I have noticed that, in this species, when the valves are in situ, what I have compared
to the wrists of the two harpes meet and close together some considerable distance
beyond and above the point of the uncus. This latter, however, is then in actual
contact with, or in very close proximity to, the stout spines that beset the ventral edges
of the harpe-arms.

X.B. — My examinations and figures arc all irorn specimens which agree minutely
with Lucas's description of " 0. rhadamantus " (Lepid. Exot. ; Paris, 1835, p. 5), called
" 0. amphrisius" in his plate ii. fig. 1, having, on the under side of the fore wing, the
nervures and nervules conspicuously bordered with light grey.

Ornpthoptera iieliacon, Pabr. (Plate XXVII. figs. 1 & 2.)

Valve more regularly rhomboidal, terminating in a finger-point well formed and sym-
metrical. Exterior scale-clothing very thin and close, therefore projecting very slightly
from the edges, mingled with a few short inconspicuous hairs. The level of the interior
forms a narrow shelf as high as the margin on each side, and then descends abruptly
to form the concavity. Lining-membrane dull, dark brown.

Harpe attached by a widely dilated base, whence it proceeds in an oblique direction,
unusually close to the dorsal margin, and there, bending abruptly at nearly a right
angle, projects more than half the length of its disk beyond the dorsal edge of the
valve — a circumstance quite abnormal. The stem is nearly parallel-sided, narrower
than usual, appearing to the eye as if a cylinder pressed nearly quite fiat, with one of
the sides wanting in the area of the disk, and at the lower half of the column. Put,
I think, this appearance is delusive, depending upon the fact that the plate of chitine
is thickened irregularly in certain parts, and somewhat bent up at the edges. Disk
(PL XXVII. fig. 2) wholly hollowed, with a very irregular cavity. The whole organ is
glittering; the uneven surface reflecting points of light from every prominence ;
colour gall-yellow, both by reflected and transmitted light. It is armed with but eight
teeth, all standing up around the edge ; but the second and fourth, on the ventral side,
are either double-pointed, or consist each of two teeth planted close together. All are
shaped as in 0. rhadamanthus. I detected no scattered bristles.

Ornithoptera heliaconoldes, Moore.* (Plate XXVII. figs. 3 & I.)
Valve almost exactly as in Haliphron, in outline ; finger-point small, but varying in
individuals ; exterior well clad with white scales, which project as a close, even fringe
of hair-scales very narrowly beyond the margins. Interior level, with a narrow shelf on
the ventral margin : lining-membrane dull chestnut, or sooty brown.

The harpe has a strong general agreement with those of Darsius and Iieliacon; yet
two examples agreed inter se in points in which they disagreed with Iieliacon. The
terminal moiety is free ; and the whole is easily detached with a fine knife-point. The
disk is merely a portion bent dorsally, at nearly a right angle, but not reaching the
margin of the valve : it does not narrow, is scarcely hollowed, scarcely thickened. It

* I give the name ou the authority of Mr. A. Doncaster, of 30 Strand, who supplied me with specimens from
Andaman Island. But it cannot be separated, even as a variety, from the preceding.

IN CERTAIN" LEPIDOPTERA. 291

resembles a very flat spoon or slice, with the edges turned-up like a dish ; there is no
sensible thickening at the bend (or " wrist "), as is usual; the teeth are ahout fourteen,
sharp and strong, varying in size inter se, standing up around the edge, and one or two
within the edge. They are not black, as usual, but of a deeper tint of gall-yellow, proper
to the chitine.The arrangement was not quite the same in the specimens ; but a common
pattern was quite discernible. The three behind the bend were much longer and
slenderer than the rest, in each example.

Uncus moderately long, little decurved, of a shining red hue, indicating, like the paler
colour in the teeth of the harpe, an inferior degree of density in the chitine ; rather
slender, the base small, the extremity subdigitate.

Scaphkim large, long ; keel abruptly deflected, much as in Rhadamanthus ; the aristate
lobe set nearer to the base.

Penis dark red, glittering ; its extremity widely ventricose, one-sided, exactly as the
organ in Haliphroh and Darsius would be if not burst. By looking steadily into the
expanded orifice, under a very good light, I could, distinctly see that there was nothing
in the throat but the wall of the organ, thick, indeed, and apparently composite, but
pellucid. Not the slightest trace of the white pulpy matter was visible in any part,

Oknithopteka Brookeana, Wallace. (Plate XXVII. figs. 5-8.)

Valve small, of peculiar form, almost circular, with irregular projections, having a
marked resemblance to that of O. Remus, but still further removed from the normal
form. There is no finger-point at the usual place ; but a prominent one at a part of
the margin diametrically opposite, viz. at the back of the dorsal margin very near the
hinge *. The exterior is densely clad with minute scales of dead-black hue, which be-
come a very wide fringe of dense brush-like hair all around the margin. The interior,
which is very concave, almost hemispherical, with no submarginal shelves, is lined with
a separable membrane, dull black.

Harpe of unwontedly large size, not only relatively, but actually ; of peculiar form,
broad, fiddle-shaped, concave, the extremity semicircular, expanded, slightly separated
from the lining membrane, and bent up ; set with about twenty-five very minute,
irregular, black teeth, along its itpturned edge.

The toothed margins of the opposing harpes converge upon the spot, within the
genital cavity, where the hook of the uncus should be. But this ancillary organ is, in
0. Brooheana, wholly aborted, the posterior edge of the dorsal arch of the eighth
segment forming — instead of a projecting triangular point, = the tegumen of Dr.
Buchanan White, — two rounded lobes, with a deep mesial notch between them (see
fig. 10) ; while the middle of the outline, which ought, normally, to be prolonged into a
curved wiry uncus, projects but a very shallow point, barely sufficient to fill the notch.

The scaphium is very large ; the two lateral hemispherical lobes well developed — with
the median furrow, along which the uncus usually passes — but no bounding crests of

* When the valves are shut, these two projecting points, one from each valve, fill up the blank left by the absence
of the usual advancing rami of the uncus.

41*

292 ME. P. II. GOSSE ON THE CLASPING-OKGANS

bristles, and no distinct "double teeth." Of the other constituent parts of this organ
I must speak with a measure of doubt. In the first specimen of this noble species that
I examined, the parts already mentioned were of a dirty drab or " whity-brown " hue;
but below them descended what I supposed to be a symmetrical keel, of pearly white
substance, pellucid, almost semitransparent ; very firm, almost cartilaginous, in texture.

But was this, indeed, the keel? For, with a very slight touch it came away, and
then seemed to be not an animal tissue at all, but amorphous .substance.

And, in another specimen of this same species, the scaphium appeared of the form
which I have with greal care delineated in figs. 7 and 8. Here the keel, orwhat I identified
as such, though of unusual shape and direction, was in the general plane of the organ,
with an open expanded extremity, hounded on each side by an ovate knob, and carrying
the usual lobes on its upper surface, without elevated crests of aristate points. Imme-
diately beneath, and even in contact with it, but in direct position, was seen the trumpet-
mouthed penis, with a horny linger-like tip.

The examination of this second specimen augmented my doubts of the nature of what
I had seen in the first ; and reminded me of what 1 have recorded in Saliphron,
Darsius, &c. But whether what I saw in Brookeana was of the same nature, and, if so,
whether it was an excretion from the penis, I have noi yet sufficient light to determine.

N.B. In fig. 8, the left valve and its harpe are indicated in faint outline, to show that
the point of the scaphium-keel came just within the serrated edges of the two closed and
opposed harpes.

Ornithoptera Amphrystjs, Cram. (Plate XXVII. tigs. 9-11.)

Valve pure white, on the golden-yellow abdomen; well-developed, large; form that of
a semicircle, with the dorsal side cut away obliquely, terminating in a strongly projecting
point. Both the margins are fringed with white hair-scales, the ventral narrowly, the
dorsal very broadly and densely, its edge forming a right angle with a point in the middle.
The finger-point of the valve is unusually large, and slightly incurved, so that, w-hen the
valves are m situ, the two points cross. I have no reason to believe that these points
have prehensile power.

On removing a valve, I could find no trace of a harpe ; and when — suspecting that
this might be accidental, the result of a morbid atrophy — I carefully examined another
specimen, in perfect condition, with well-formed abdomen, I still found nothing to break
the uniformly concave surface of the valve. Only, at the base, near the dorsal side, there
was, in each, a pear-shaped area, occupied by a multitude of very fine and close corruga-
tions in parallel lines. But, afterward, when I was examining the uncus and its accom-
paniments, the mystery was solved. I observed a long black slender wire, incurved at
the tip, projecting from the interior, nearly horizontally, below the scaphium; which I could
not at all account for. Tracing this to its origin, I found that it belonged to that one of
the valves that yet remained in situ ; that the place of the corrugated area was normally
occupied by a curved plate of polished brown chitine of trapezoidal figure, from the dorso-
basal angle of which proceeded the long curved wire, of extreme tenuity throughout.

IN CEETAIN LEPIDOPTERA. 293

Tlie two inner angles of the trapezoid were stout short points, rising free from the valve,
and all of deep hue and glittering surface, indicative of great density.

This, then, is doubtless the harpe ; and the reason why I had not found it in either of
the separated valves is, that it is very readily detached from the lining-membrane; and
so, when the valve is removed from the abdomen, it is apt to leave behind the harpe,
which remains in situ. It is, indeed, widely different, both in form and position, from the
organ, as I have found it in all the other species of Ornithoptera that I have examined ;
but it is probably very effective, as I hope to show.

The uncus is long, slender, wiry, horizontal, nearly straight, bent slightly down at the
tip, proceeding from a very small triangular base. The scaphium is greatly developed,
forming various folds, the usual white hue becoming brown, and even black in several
places. Its " double tooth " is a long horizontal spine on each side of the uncus, with
only a slight projection behind it ; and, behind this, the upper edge of the scaphium l'ises
above the level of the uncus, in several minute bristled points.

The penis is thick, enlarging to the extremity, which is truncate, with a corrugate
expansion, and a terminal finger-point, which, as well as the denser, darker side of tin-
column, is here placed below. It issues from a sheath formed by a descending fold of
the scaphial tissues. Between it and the scaphium on each side projects the long w
of the harpe already alluded to. Though its position and direction, when in the valve,
look awkward, yet, when viewed in situ, we see that, on each side, this acute elastic wire
brings its curved point just below the spot where the similar wire of the uncus comes ; so
that the three points converge, and doubtless secure a very effective grasp of the female
abdomen ; though I must confess the scaphium appears to be very much in the way.

The 0. amphrisius mentioned by Dr. De ITaan (see p. 267, supra) is certainly not this
species; but appears to be the Amphrisius figured by Lucas (Lep. Exot. ; pi. ii. fig. 1),
which is the same as the preceding O. rhadanianthus of Boisduval (ut supra, p. 289).

Ornithoptera Remits, Fabr. (Plate XXVII. figs. 12-20.)

Voice transversely ovate ; a small horny curved finger-point on the dorsal corner of
the front edge, not quite constant, being sometimes reduced to a low prominence. There
is another similar prominence somewhat more in front, which also varies in height.
Little fringe is seen, and that only on the dorsal margin. The cavity is diminished by
an obscure, wide, flat shelf, extending along the ventral slope, studded with minute
bristles pointing inwards, and also by a curious large knob, at the dorso-anterior corner,
very constant, which is followed, in front, very abruptly, by a deep sulcus. This knob
is black, hirsute, often hoary, and lies immediately behind the finger. Occasionally, the
valve, when detached, brings away a holloAV, shell-like, triangular plate of chitine, which
had formed part of its hinge.

Harpe an irregular rotundo-triangular shell, or ear-like body, of polished black chitine,

broad, but unusually short, often not reaching halfway from the base to the point of the

( valve. Its edge is seen to be minutely serrate. When removed, and viewed on the stage

of the microscope, very obliquely, so as to bring the free edge (the ventro-anterior edge)

291 ME. P. H. GOSSE ON THE CLASPING-OKGAN>

into full square view (as in PL XXVII. fig. 13), the serration of the edge is seen to be
very fine, and slightly irregular. Its base has a remarkable upturned lobe.

Tbe form of this organ may be well imitated, and illustrated, by cutting a paper
pattern into the shape of fig. II, and rolling it into a tube, of which the dotted line is the
longitudinal direction. If then it be pulled gently open, it will assume the exact curva-
ture of the harpe. The long oblique point at the lower end of fig. 14 is, in situ, upcurved,
and appressed to one of the flattened knobs which occupy the floor of the abdominal
cavity. The expanse of the harpe resembles clear yellow glass, intensified into nearly
black at the edges all round.

The uncus is of regular curvature, about one fourth of a circle, moderately slender,
black, polished, coming down just between the two shell-like harpes. It is closely em-
braced by the lateral lobes of the scaphium, in which I sec no bristle-bearing crest, and
no " double teeth." But the upper edges arc pilose, with short, stiff, close, shining hair,
pointing towards the dorsal surface of the abdomen, as is usually the case with those
species of Ornitlwptera which have not aristatc crests.

The penis is a moderately thick red-brown column, which protrudes from the midst of I
the eighth segment into the midst of the genital cavity, curving upward, nearly to tin1
tip of the uncus. It is obliquely truncate at the point, where (in the specimen cxainined)
there was what had the appearance of a glans, surrounded by a preputial thickening.
But this is, indeed, so much like the remarkable appearance described and figured in
O. Haliphron (see p. 285 supra) that I am inclined to refer it to the same origin — as if the
pulpy substance that (at times) occupies the interior of the penis had been forcibly pushed
out from within, against the underside of the scaphium. Here, however, the walls of the
penis had not been burst, as there ; the orifice was not sensibly trumpet-mouthed ; and
there was no other trace discernible of the pulpy matter. The mass in question, more-
over, instead of being white, was dull brown. Possibly, these features may indicate that
the extrusion had not been very recent. The organ is reversed.

The two projecting knobs of hard polished chitine, which serve as ancillary supports
for the bases of the valves and harpes, are large and conspicuous. The use of the knobs
admitted of no doubt, as I, slowly and gradually, removed one of the valves *.

Papilio Memnon, Linn. (Plate XXVIII. figs. 1-3.) [Cram. i. 91. c— Borneo.]

T'alce-outlme ovate, obliquely truncate at base ; edge narrowly turned-in, or forming a
narrow shelf fringed with short spare hairs ; interior floor smooth, but not polished.

Harpe, a narrow blade of dark polished chitine, which, springing from the base of the!
valve, runs close to the edge of the shelf, all along the ventral margin, to the very point, |
where it ends abruptly. This long blade may be compared to a knife affixed by its back
to the lining membrane of the valve, but sloping inward, its free edge having several
excavations in its outline ; the organic adhesion of the lower edge is, moreover, lacking
for the terminal fourth of its course.

This blade can, without much difficulty, be lifted out of the valve, the basal mem-
brane yielding to force. The outline now can be more correctly perceived. At PL XXVIII.

* For further notes on this species see Appendix.

IN CEETAIN LEPIDOPTERA. 295

fig. 2, the longitudinal portion is shown, laterally, more magnified, and viewed as a
transparent object on the stage of the microscope. It now appears still more like a
knife-blade, of thin material, transparent, but of a deep yellow hue, the surface studded
with numerous minute round eminences, each of which is surmounted by an erect
bristle. The inferior, adhering edge, is thickened into a narrow rim, which is fringed
with a row of close-set short hairs which grow in a direction across the blade. The free
edge is notched into a number of minute saw-teeth, not quite uniform in size or distance.
These serrations are seen only on the blade proper, the free terminal fourth.

The abdominal armature is normal, the miens, the scaphium, and the penis, being all
well developed and of ordinary proportions. When the valves are in place and closed,
the tips of the two harpe-blades meet close to the tip of the uncus. The scaphial
" double teeth " have this peculiarity: the inner (base-ward) member is obsolescent ; the
outer stands up quite perpendicular, a straight needle of polished horn. The penis has
a broad expanded hiant extremity, which turns towards the right side. Fig. 3 shows
these parts in situ, the right valve having been detached, the left and its harpe indicated
in outline.

The continental form, marked by broad red ocelli with black pupils, more or less con-
spicuous, at the anal angle of the hind wings beneath, has been lately distinguished,
even by such an experienced naturalist as Mr. Wallace, as a true species (Tr. Linn. Soc.
xxv. p. 47) ; and Mr. Kirby gives it a specific place as P. Agenor (Syn. Cat. 552). But
if any faith is due to the organs of which I am treating, the distinction cannot be
maintained.

I examined a <$ , which agreed, in all essentials of form and marking, with Cramer's
pi. xci. A, b, whose name of P. Androgeos Mr. Wallace would retain. But I can find no
feature of valve, harpe, uncus, scaphium or penis, in which it differs to any appreciable
degree from examples of P. Memnon received from Borneo.

Papilio Mayo, Moore. (A. D.) (Plate XXVIII. figs. 4-8.) *

Valve agrees closely with that of P. Polymnestor. So also does the harpe, as it lies in
situ; perhaps still more with that of P. Memnon. But when lifted from the valve (which
is done with readiness) and examined separately, there is considerable difference. The
blade is not so regularly securiform, and the end is not so decidedly bent-over ; the
blade is rudely triangular, serrated with minute teeth all along its dorsal edge, and the
greater portion of its broad end.

The uncus and its accompaniments are on the type which I have figured in P. Memnon.
The scaphium is very well developed ; and I take advantage of this to make a more
careful examination of the organ.

If we compare the uncus to the upper maudible of a bird's head, Ave see a white fleshy
body filling the place, and taking much of the form, of the lower mandible. In this

* Tor a few species examined, I have been dependent on loose abdomens, supplied by Mr. Arthur Doncaster, of
30 Strand, London, who has verified their names. These are indicated by the initials above.

296 ME. P. H. GOSSE OX THE CLASPIXG-ORGANS

species, as in many others, it takes a boat-shape, swelling into rounded sides, and nar-
rowing abruptly in front.

If now with the point of a needle, we raise the upper mandible till we break it off by
force, we see more clearly the upper surface and attachment of the scaphium. At the
back of the " palate," there was an organic attachment of the scaphium, all across its
width (see figs. 7 and 8), apparently perforate in the middle ; and this perforation I take to
be the anal orifice, though I have not been able to trace the connexion of it with the
intestine. The middle and front are hollow and capacious ; the edges (the " gunwales" of
the boat) are high, broad, and flat ; their summits crowned with what mimic true
molar teeth. In the present species, the apparatus (on each side) consists of, first, a
small acute, recurved tooth, with a secondary one by its side, nearly transverse to the
(supposed) jaw (fig. 7), them a low conical point, ami behind this a flat black mass, like
a real grinder. The verisimilitude of these points to real teeth, and the organ which
carries them to a real jaw, is wonderfully exact, especially as we cannot suppose any
chewing-function to exist.

The back {apparently, but really fore) portion of the scaphium sends off, below, two
rami of the same white flesh, which are soldered to the sides of the eighth segment, near
the bottom, leaving a slender aperture through which the penis protrudes, whose sheath
is, I think, made by, or united with, these rami (fig. 6). This organ comes forth
pointing upwards, but then bends to a nearly right angle, and points obliquely down-
wards. It is flattened, and even guttered along its upper surface, and ends in a bilid
point, dilating below into a broad membranous expausion, whose sides fold together, the
inferior half of the column apparently the more densely chitinous.

The segment ends below in two large, blunt, extremely polished, hollow tooth-like
knobs, placed close together, the inner supports of the harpes.

Papilio Euechtheus, Donov. (Plate XXVIII. figs. 9-13.)

Valce-ovitlmc an oblique triangle with swelling sides, or a pointed semi-oval ; both
margins fringed, the dorsal chiefly, the fringe running to an acute point : interior surface
grey, dully shining ; a narrow, deep channel runs through the ventral half of the cavity,
which else is shallow.

Harpe a knife-like ridge of rich brown chitinc, highly polished, affixed by one edge to the
bottom of the channel, and running through the length of the valve, parallel to the
ventral margin. It has a narrowly dilated base, a wide triangular tooth projecting from
near the middle and leaning-over dorsally, and a wide hatchet-like head, which also leans-
over. Prom the adherent edge, on the dorsal side, springs a line of close-set bristles
having a golden lustre.

When the harpe is removed from the valve, and examined by transmitted light, both
the median tooth and the head appear as thin plates of clear, sienna-brown glass, each
brought to a keen edge, and cut into minute saw-teeth, not perfectly uniform. Both
plates show dark irregular lines of corrugation meandering like map-rivers on their
surfaces.

The uncus (figs. 11, 12) is the beak-like produced extremity of the eighth segment, which

IN CERTAIN LEPLDOPTEEA 297

takes conspicuously the form of a bird's skull. Ordinarily this part is clothed densely
with the common scales and hair-scales ; but, when denuded of these, it is seen to he
smooth (not polished), thin, pellucid, light brown, composed of a very firm and bony
chitine. There is no joint between the ventricose " skull" and the slender "beak," so
that the latter can move only as the whole segment moves ; but its acute point (as seen
in figs. 12 and 13) falls normally just between the two terminal plates of the harpes when
the valves are closed. The uncus is slender, wiry, dark brown, glittering, moderately
curved, its tip slightly spathulate, its rami coming to a point horizontally and verti-
cally at some distance short of the tip.

ScapMum essentially as in P. Mayo ; but the sides are much less bulging, the keel not
pointed, and the black knobs behind the erect "teeth " are replaced by tufts of crowded
brown bristles, which much resemble those usually seen in the Ornithopterce, but are not,
as there, limited to a linear series on each side. What in Mayo I compared to the gun-
wales of a boat, are here flat triangular expansions.

The penis is short, with a thin expanded orifice, shrivelled, showing no apparent
structure.

Papilio Rhetenob, Westwood. (Plate XXVIII. figs. 11-16.)

Valve of an outline resembling that of the outer half of a lemon, oblique, the ventral
edge bv much the longer ; exteriorly dull black, thinly clad with scales ; hair-scales
projecting in a wide and close fringe beyond the dorsal margin, augmented by a forest
of slender bristles springing from the interior, so that the actual edge of the valve
is here much concealed, except in oblique lights. Along the ventral margin the fringe
is thin, and composed almost wholly of bristles from the cavity.

The harpe, viewed in situ, PI. XXVIII. fig. 11, is a long tubular rod of glittering
chitine, terminating in an oblique hatchet-shaped head, facing the dorsal side, and set
with strong bristles below. On carefully removing it from the valve, and changing the
angle of vision, we find the head to be much of the form of an American woodman's axe,
fig. 15. Its edge is of semioval outline ; and, under increased magnifying power, is seen
to be notched with teeth of excessive minuteness and of no great regularity, of which the
three combined lenses of a specially powerful pocket-glass revealed no trace.

The uncus is nearly bill-shaped, with slight curvature, sharp-pointed, deep-bladed, the
summit crowned with close, tall, erect hair-scales ; the rami stout and broad.

ScapMum large, with a deep vertically truncate keel, wide margins, on which
stand conspicuous " double teeth," the principal one of each pair being a tall recurved
spine. The white fleshy rami of the keel descend to form an investing sheath, whence
the penis protrudes to a great length, almost horizontally, yet in a sigmoid curve, ter-
minating in a finger-point above, and a triangular expansion with thin edges, below.

Papilio Deiphontes, Wall. (Plate XXVIII. tigs. 17-19.)

Valve of the same form as in P. Pihetenor, Proienor, &c, but peculiar in having the
cavity marked on both sides, towards the point, with a number of parallel ridges ; both
margins are broadly fringed.

The harpe, in situ, bears a close resemblance to that of P. Polymnestor, a slender

SECOND SERIES. — ZOOLOGY, VOL. II. 42

298 ME. P. H. GOSSE ON THE CLASPING-OEGANS

tubular rod, with a hatchet-shaped tip heut-over dorsally. But, separating it from the
membrane of the valve (to which it adheres somewhat tenaciously), we find that the
end, which stands obliquely erect from the plane of the valve, is broadly semicircular
in outline, and that this outer edge is serrated with delicately formed but sufficiently
strong spine-like, acute teeth (as seen in PI. XXVIII. fig. 18.).

The uncus is long, slender, wiry, curved to one fourth of a circle, clothed along the
median line of its summit with a ridge of erect black hairs, tall at the base and dimi-
nishing to nothing near the point ; the point a small blunt spathula ; the under surface
a narrow keel of uniform depth.

Scaphium little developed in its white fleshy portion, but well in its horny arma-
ture; two black, shining, upcurvcd, stout teeth projecting from the extremity, each
with a low conical knob behind it. Between these, the usual fleshy keel is represented
by a narrow plate of shining brown chitine ; and true keel there is none*.

The penis, of ordinary form and size, projects from the very bottom of the cavity,
leaving a great vacant space between it and the scaphium.

Papilio Schmeltzi, Herr.-Schaff. (A. D.) (Plate XXVIII. figs. 20-22.)

Voire rather small, pointed ; the narrow hair-like black clothing-scales of the
exterior projecting beyond the margins slightly and raggedly. Cavity shallow.

llarpc a chitinous ridge (or thin blade set edgewise) of rufous hue, which runs through
the length of the valve (PI. XXVIII. fig. 20, right valve) in a sigmoid line, not
separable from it without tearing. It does not appear to dilate into a lining plate
throughout its length. The base spreads like that of a tree-trunk. At the lower curve
of the ridge is projected a stout triangular tooth ; and the extremity dilates into a broad
axe-shaped blade, somewhat thick, as if the axe were unfamiliar with the grindstone.
The dorsal side of the ridge is fringed with a line of stiff hairs.

In these observations I have several times fancied a slight difference in the armature
of the two valves of the same individual t- In this example it was quite manifest; for
in the left valve (fig. 21), the sigmoid curve was very much slighter; the lower knob
was more produced, and more decidedly a tooth, bent-over dorsally; and the securiform
head differently shaped, itself approaching the form of a broad tooth ; both projecting
towards the dorsal side at an angle of about 15° with the floor of the cavity. There is,
too, a secondary ridge, very evanescent, running down from the point toward the dorsal
margin of the valve; and a third ridge, which seems not connected with the harpe,
running through the length of the valve, centrally.

Up to this point my observation had left the impression that the machinery seemed
inefficient in this species; the rounded outline and blunt edge of the "axe-head"
seemed incapable of any seizing or holding function, like that of the sharp points with
which I had become familiar in these elaborate harpes. But, on submitting it to an
increased magnifying power, viz. of 140 diameters, the " axe " assumed a quite different
aspect. Now the edge was seen to be studded closely (see fig. 22) with minute, acute,

* But this needs reexamination.

t See the Memoir by Messrs. Scudder and Burgess, cm Asymmetry in the Hesperiada1, in Proc. Boston Hoc.
Nat. Hist, for April, 1870.

IN CERTAIN LEPIDOPTEKA. 299

curved teeth, which were also distributed for some distance from the edge, ou the surface
of the blade. Thus a resemblance is suggested with the harpe in Omithoptera of the
Priainus group, whose crowded glittering points make, doubtless, a sufficiently effective
prehensile apparatus.

The uncus does not differ importantly from the same organ in P. Erechtheus. It is
deeply channelled along its upper surface ; and the channel, as well as the summit-ridge of
dense black scales, very long and narrow, extends almost to the spathulate point. The
white fleshy conical scaphium with the " double tooth " on each side, is as in P. Bemoleus.

Papilio Polymnestor, Cram. (Plate XXVII. figs. 23, 24)

Valve long and narrow, blunt-pointed, semiovate, deeply concave ; the margin sharp,
not overarching in any part. The dull black scales of the exterior project ruggedly all
round to a slight extent, and there is a crowd of slender hairs growing from the floor
of the cavity, especially near the tip and the dorsal margin, most of which reach beyond
the edge, but do not constitute a fringe.

The harpe is of the securiform type, yet with peculiarities of its own. The hollow
between the head and the secondary tooth is here filled up, so as to form a long
blade with a continuous even outline, which follows the curve of the ventral margin of
the valve, and leans considerably inward. The very edge is, so to speak, bevelled off ;
and, at the lower end, where would be the secondary tooth in the most usual form,
project from the very edge six (or seven, one being double) exceedingly minute saw-
like teeth. These, as well as the whole plate, are highly polished ; indeed, the glittering
brightness and deep colour of these prehensile structures, generally, are most attractive
features, and convey the idea of their almost metallic hardness. High magnifying
power showed minute grey body-scales, still adhering to these teeth, and to the extreme
point of the harpe — trophies, doubtless, of battles fought, proelia Veneris, in which
these are effective weapons *.

Papilio Protexor, Cram. (Plate XXVIII. figs. 25, 26.)

The valve is of the ordinary form, margined by a broad fringe of hair-scales along the
dorsal, but nearly bare on the ventral side. The latter has a flat ledge or shelf bordering
the margin within, whence the cavity abruptly descends. A broad blade-like harpe
pursues a corkscrew-like course, affixed by one edge, and running to a free point, the
terminal half of the upper edge serrated with numerous, tall, close-set, curved teeth,
even to the tip. PI. XXVIII. fig. 25, shows the interior of the right valve with the
harpe in situ ; fig. 26, the harpe separated, viewed from the ventral side, and further
magnified.

The abdominal armature agrees generally with that of P. Phetenorand Erechtheus ; the
uncus and penis with the former specially.

* Professor Owen says (Comp. Anat. Invert. 1855, p. 397): — "The structure of the intromittent organ in the
Lepidoptera is such as to preclude the repetition of the art ; ami they consequently live in a state of compulsory
monogamy. The bifid hooks on the terminal segment of tie.' dorsal valve of the penis, whilst they seem to retain the
female, prevent the entire extraction of the virile organ."

I cannot reconcile this description with anything I have seen in the Papilionidse ; and I strongly doubt the
inference.

42*

800 ME. P. II. GOSSE ON THE CLASPING-ORGANS

Papilio Ascalaphus, Boisd. (Plate XXVIII. figs. 27, 28.)

Valve having the outline of the terminal half of a lemon cut obliquely. Interior
cavity very slightly fringed on the dorsal side alone, and that chiefly with hairs growing
from within ; umber brown with a dull satiny gloss. Harpe of the pattern of the
Memnon group, near to that of P. Dcij>hoi//cs; and, like that, serrated along the free
erect edge, the form, however, differing in detail, as shown in the figures (27, in situ
in the valve ; 28, removed, viewed from the ventral side, and further magnified).

Uncus with a similar curve to that of P. Beiphontes, similarly clothed above, and
similarly keeled ; but not nearly so wiry, widening rapidly from the point. The
scaphivm is furnished with a thick white keel, whose descending rami behind — that
combine to embrace the penis, as a sheath for its basal portion — are so short that this
latter organ protrudes obliquely upward, its extremity appearing almost close beneath
the scaphium-keel*. The " double tooth " is very small, but distinct, erect, and curved
back.

I have not yet found any two species so closely assimilated in the details of the
prehensile apparatus as Beiphontes and Ascalaphus; yet the differences here, if mostly
minute, are quite distinct. The radiating ridges of the valve, the marginal fringes,
the axe-head of the harpe, the scaphial keel, the scaphial double teeth, and the position
and direction of the penis, all present marked diversities of form and detail.

I subsequently examined another specimen of Ascalaphus. It agreed with this one
in the minutest details, except that the tiny half-separated lobe at the point of the
harpe I could not find. This was, therefore, in all probability, accidental.

Papilio Hllenus, Linn. (Plate XXIX. figs. 1-3.)

We have in the valve of this species almost the very counterpart of Ascalaphus. The
harpe, in situ, looks different, indeed, because the extremity is not bent-over quite so
much, and the line of direction is more incurved in the middle. But, when it is viewed
separately, and from the side, it takes almost precisely the same form, broadly cres-
ecntic ; and its edge is, in like manner, serrated with excessively minute spines.

So the abdominal apparatus is of the same type. The uncus, circularly curved, is ridged
above with still' hair. The scaphium is almost exactly on the pattern of P. Demoleits ; but
the double tooth more developed, high, slender, acute, incurved. Finally, the penis
has its more chitinous side uppermost, is dilated at the point, and is of considerable
thickness and length.

Papilio Nephelus, Boisd., differs in nothing of importance. The harpe runs very
close to the ventral margin, but its form is exactly that of P. Ascalaphus. Scaphium
large, and its double teeth almost exactly as in P. Hclenus.

Papilio Chaou, Westw., scarcely differs, but the harpe is not quite so close to the
ventral margin.

■ Subsequent observations lead me to doubt whether the direction of this organ is a character of specific value —
whether it is not one of the conditions which vary at different times, even in the same individual.

IN CERTAIN LEP1D0PTEEA. 301

Papilio Menestheus, Drury. (Plate XXIX. figs. 4-6.)

Valve-outime a long, narrow, pointed arch, swelling on the dorsal, excavated on the
ventral side ; fringed with black bristly hairs around the point, which grow from within
the margin, and disguise its form ; the ventral margin, from the excavation downward,
fringed with close-set long white hair-scales, connected with the clothing of the exterior.

The harpe, a slender curved rod of polished black chitine, running up close to the
ventral margin, and following its outline; it sends off a strong, acute, conical spine
about the middle, which is free, pointing to the dorsal base; in its terminal half, it
expands into an ovate blade, convex ventrally, concave dorsally, standing on one edge,
and having the free edge, as well as the spine, minutely serrate. PI. XXIX. fig. 4
represents the right valve, with the harpe in situ ; fig. 5, the harpe dissevered from the
valve, viewed ou the stage of the microscope as a transparent object. It is now wholly
of a rich sienna-brown ; the edge cut into curved spinous teeth, far more minute and
far more numerous than I have been able to figure. The side, moreover, is marked
with a multitude of fine slanting lines crowded together, which, by delicate manipu-
lation of the magnifying power, resolve themselves into so many linear series of very
minute pits on the surface, the edge of every one reflecting a semicircle of light.

The abdominal apparatus has much in common with that in P. Beiphontes, &c. ; and
the valvular armature may be resolved into the same type (fig. 6).

Papilio Pammox, Linn. (Plate XXIX. figs. 7-9.)

The valve and the harpe belong to that genera] type which we see in the Memnon
group, the former presenting little notable, the latter coming near to the forms in Deiphontes
and Ascalaphus, especially when it is detached from the valve, as at PL XXIX. fig. 8;
for, strange as it may seem, the outline of the harpe, when detached, is not always the
same as when it was in situ. In this case, the disk, which takes a hatchet-form, occupies
a much greater area of the valve (fig. 7) than in the species named ; it is, as in them, of
polished, pellucid, brown chitine, the edge serrated, rather unevenly, the whole stem (the
haft) and a broad area at the back of the blade (shown at tig. 8) having been in contact
with the floor of the valve, while the blade itself stood up free.

The uncus is slender, curved almost to a semicircle, with a small spathulatc tip, a
moderate vertical divaricate beam below, notched on each ramus, and the ridge crowned
with erect hairs (fig. 9), which do not diminish in height to the tip, but end abruptly.

The scaphium is Avell developed, nearly free from lateral folds; the "double teeth"
prominent, the principal one of the pair projecting obliquely towards the tip of the uncus :
the fleshy keel thin, short, but very deep, ending in a minute point. The jJCnis is long,
moderately thick, truncate, scarcely expanded.

At the bottom of the abdominal area (fig. 9) is shown the nearer of the pair of pro-
minent angular knobs, to which had been attached the triangular base of the harpe-
shaft, and that of the right valve.

Papilio Agaves, Drury. (Plate XXIX. figs. 10-13.)

Prom a certain resemblance in the form and colours of this South- American butterfly
to the Indian P. Sector — near which it is placed, in efforts made by entomologists to

302 ME. P. H, GOSSE ON THE CLA8PING-ORGANS

arrange this vast genus in naturally allied groups — J expected to have here another
example of abnormality, such as we see in the Polydorus group. But, except that the
uncus is abruptly bent perpendicularly downwards, as it is in P. Hector, the organs are
of the usual type, though peculiar in detail.

The valve is small, semioval, the outline slightly bulging dorsally and fringed only with
a few very minute hairs. The interior is almost level for a wide space within the ventral
margin, then abruptly descends with a deep cavity, which extends close to the dorsal
margin ; internal surface cloth-like, nearly black. Within this cavity lies a great scmi-
crescentic harpe of shining chitine, its lower half divided off into a kind of stem, the whole
somewhat like a butcher's cleaver with its handle, and all springing from a dilated base by
which it is affixed to the proper knob, within the bottom of the 8th segment, apparently
distinct from, though in close proximity to, the conjunction of the valve itself with the
segment. By removing the harpe, which is very readily separated, we see clearly (what
is, indeed, discerned by the aid of a pocket-lens, even in situ) that the straight side, or
what answers to the cutting-edge of the cleaver, is serrated with sharp teeth, some-
what irregular both in size and arrangement ,*.

The abdominal organs (fig. 13) are remarkable for the abrupt deflection of the uncus
(perhaps accidental), the extreme minuteness of the scaphium, and the great development
of the penis, which is reversed.

Papilio Machaox, Linn. (Plate XXIX. figs. 14-16.)

In our own familiar Swallow-tail the calve is of the ordinary subtrigouate outline,
with a rounded point. Externally it is clad with a close dense coat of pale-yellow scales,
which do not project beyond either margin ; there arc, however, black hairs, growing
from the interior, more than usually stout, long, and close, along the dorsal side, and
forming a produced point. These crowded hairs, the whole cavity being of a sooty
brown hue, effectually conceal the marginal line on that side.

A low pale ridge runs through the valve along its central line, which does not seem
connected with the harpe. Just within the ventral margin, following its line from base
to tip, runs a plate of glittering black chitine, fixed to the floor by one edge, while the
other is free, and leans over towards the dorsal side. At mid-length its height abruptly
increases ; and thence to the point it is a keen knife-blade, only cut into minute sharp
teeth, about a score in number, exactly like a saw. This harpe is of chitine so dense as
to appear quite black, but transmitting a rich translucent golden-brown hue when
viewed against the light, except a slender line just below the teeth, which is considerably
paler, as if there were a band of less density just there, or possibly a groove, though
it is not a place at which one would expect the structure to be weakened.

The shining chitinous material is spread widely over the valve-cavity as an exceedingly

As a rule, I content myself with delineating one of the valves. Bui of P. Ar/avus I have given the left as well
as the right. In the course of examination this left valve came away, leaving the harpe still attached to the segment.
When 1 detached this, in turn. 1 observed what appeared an accessory piece behind the blade, and following a similar
curve; this I delineated as 1 saw it, without understanding it (see fig. 12.) Eeplacing the harpe in the valve,
I drew- it, as I supposed, in situ, as at fig. 11 ; but, afterwards, having detached the right valve and its harpe entire,
I saw nothing of this additional piece ; I therefore conclude that it is a piece torn off accidentally from the back of
the harpe.

IN CERTAIN LEPIDOPTEKA. 303

attenuated plate. It adheres so tenaciously, that I did not succeed in separating it from
the lining-membrane without much laceration.

Another example agreed minutely with this description. The only differences that I
could detect were, that the valve was slightly narrower, and more acutely pointed, the
outline being a nearly perfect isosceles triangle. The harpe extended to the very point of
the valve ; yet along its edge I could count only fourteen teeth.

The abdominal armature, in general, agrees with such examples as P. Ilayo. The
uncus (in the specimen figured) Avas long, narrow, the margins much elevated, displaying
a deep keel, and was so strongly downcurved that the small, polished, spathulate tip was
bent considerably within the perpendicular ■'. Its upper surface is partially concealed
under a dense thatch of long arching hair, mingled black and yellow, which, springing
from its basal part, projects far beyond its doAvnward curve f .

The scaphium is of moderate dimensions, and of normal form, with a rather small
keel. The lateral expansions, " gunwales," are wide, and are furnished with " double
teeth," the principal of each pair being sufficiently conspicuous, projecting almost
horizontally. There is a strong projecting point from each lateral edge of the tegumen.

The penis is moderately large, trumpet-mouthed, the expanded lips notched (corru-
gated ?) along the edges, and surmounted by a short finger-point. Its protrusion from
the abdomen is high, even so as sometimes to press the under side of the scaphium.

Fig. 16, the abdominal organs viewed from the right side with the right valve removed ;
the left in situ, shown in outline.

Papilio Akcttjrus, Westw. (A. D.) (Plate XXIX. figs. 17-19).

Fefoe-outline that of a Gothic arch, with the point rounded and the dorsal side a little
hollowed ; a wide fringe of coarse brown hair-scales along the dorsal margin. The cavity-
floor has an abrupt depression along its middle, broad at the base, and tapering to the
point.

Harpe a slender wire bounding this depression ventrally, rising at about two thirds of
its length into a rather broad blade, the edge of which leans-over dorsally. More highly
magnified, it is seen to be somewhat semiovate in outline, curved at the tip into a strong
hook, and beset, along its free edge, with a number (about forty-five) of minute close-set,
sharp saw-teeth, the intervals between which run down in grooves upon the blade ; in
other words, the teeth are the extremities of ridges. The whole harpe can be separated
from the valve only by tearing the adjacent portions of the lining-membrane, to which
it appears to adhere organically, for almost its whole length. Some examples have the
blade less irregular in form.

The abdominal organs are almost the counterpart of those in P. Mayo. The uncus
presents nothing worthy of remark. The scaphium is scarcely distinguishable from

* But examples differ in this respect, the uncus being, in some, nearly horizontal at first, and abruptly and strongly
curving toward the tip.

t The organ is perhaps unusually subject to malformation ; for, in one example, its form and curve were quite
asymmetrical, and it was accompanied by a supernumerary tooth, much like itself, but shorter, on its left side. Yet
another example had a shapeless knob projecting from the right side of the tegumen.

301 MR. P. IT. GOSSE ON THE CLASPING-OPGANS

what I have described and figured under that species, save that: — 1. its keel has no
minute point, but ends with a deep round thin edge ; 2. the double tooth, instead of
being followed by a flat black knob, has, in place of it, a second minute conical point,
and clusters of bristles.

On my lifting the uncus till I broke it from the tegurnen, I could see, by means of a
lens, a distinct orifice. I studied it carefully, and am sure it was not a rent, but a true
organic orifice, tub alar, central in position, under the tegumen, with minute protrusile
lips, which merged into the sides of the scaphial groove. This surely was the anus.

The keel, or " cutwater," runs olf, as I have just said, to an extremely thin vertical ,
plate. This I suspect to be the seat of some special function ; for the terminal portion,
for a considerable width from the edge, well-defined, lias a peculiar surface, finely
granular, almost pilose, whereas the other parts of the surface are shining, white, smooth,
though plicate.

I am again struck with the wonderfully close resemblance which the scaphium and J
the uncus together bear to a bird's skull and beak, with a mammalian lower jaw.

The penis is of moderate size, pointing obliquely upward, with expanded lip, thin,
transparent, corrugated. It proceeds from the midst of a white fleshy veil *, which oc-
cupies the entire back of the genital cavity, formed by the dilated rami of the scaphiumf.

Papilio Merope, Cram. (A. D. parthn.) (Plate XXIX. tigs. 20-23.)

Valve subtrigonal, more or less regular ; furnished with a fringe of thick, close, and
fine hair-scales, of pale buff hue, which widens uniformly along each margin from
base to point. Internal cavity dark brown, smooth but not polished, and beset with a
few fine hairs.

Harpe a thin and very narrow lamina running through the whole length, near the
ventral margin, and subparallel to it, affixed by its inferior edge, and leaning-over
towards the dorsal side. It terminates sometimes in a fine-drawn point, which even
extends beyond the limits of the valve, as far as the point of the fringe. At other times
the blade is rounded at the end, and of equal width throughout from its abrupt expan-
sion a little way above its base. The terminal third is free, and springs up at an
acute angle from the lining-membrane. The whole blade is serrated with minute teeth,
more than fifty in all, pointing backward. From the point where the stem dilates to
form the blade, a curious finger-like process is given out, which goes nearly across the
valve dorsally ; it is straight, cylindrical, obtuse, and unarmed.

The uncus makes a curve of about a quadrant of a circle, ending in a very sharp spa

* Equivalent (o the " curtain " of the Supplementary Note to Orniihoptera /A ////'.•-. p. 336, infra.

t By a subsequent observation of 1'. Arcturus, made under unusual advantages of sun-light, I am able to supple-
ment (he above description, but unfortunately not in time to add some figures to Plate XXIX. The scaphium has
little resemblance to that of /'. Machaon or Zolicaon. It is a perfect boat in form, of what a sailor would call beautiful
"lines;" apparently quite empty, save for the two great round-bottomed feeth, one on each side, which take the '
exact form of the boat-curve. The "cutwater" runs off in front straight and clean, its front edge sloping away
downwards ; though it is thin, it is immediately behind abruptly hollowed to a yet thinner lamina. The " gunwales "
of the boat expand amidship, much like those of a paddle-wheel .steamer (when viewed vertically from above), and
then descend to unite with the wall of the abdominal cavity. The great scaphial teeth arc stout upeurved spines,
studded sparsely -with minute hairs, while ridges behind and between them bear long, stout, erect bristles.

IN CERTAIN LEPIDOPTERA. 305

tlmlate point, and a deep angular keel. Its summit is crowned with an erect ridge of
stiff hair, pointing slightly forward (fig. 23).

The scaphium is more than ordinarily developed, and it presents some peculiarities
so suggestive that I bestowed upon it more than usual attention. The mass of flesh, of
conspicuous whiteness, and glistening at all its prominences, occupies almost one third of
the volume of the visible cavity. The keel is deep and ample, though thin ; it is much
excavated, the surfaces full of irregular hollows, with slender threads, and isthmuses
of white connecting flesh stretching across ; the hollows are more or less tinged with
black, and of a minutely granular surface.

The " double teeth" attain here a development to which I have observed no parallel*.
They are seated in the usual place, on the summit of a great rounded mass of polished
flesh, which swells out on each side (fig. 22) from beneath the base of the tegumen,
gradually becoming of a deeper and deeper hue, apparently more and more chitinous,
until the colour merges into the glistening black of the teeth. Of these the principal
— which one can scarcely help calling the anterior (though it is strictly the posterior),
being next to the point of the uncus — is a broad, curved, acute spoon, cut on both its
edges, and also on Us bowl, with stout serne. The other tooth, equally tall, but
slenderer, curved in the same direction, viz. toward the base, is similarly serrated on
several lines. The whole space between the two teeth is a continuation of the spoon-
bowl, and is everywhere studded with sharp points. It is strange that most, if not all,
of the serrations are surmounted each by a stiff golden bristle, which seems likely to
interfere with their effectiveness f ; but the number of pale-yellow body-scales clinging
to all parts of these notched teeth show that they have done duty as prehensors.
Certainly, viewed in sunlight under the compound microscope, they are formidable
weapons to look at.

At its very base, where it springs from under the deep keel of the uncus, the scaphium
forms a globose knob, skull-shaped, which is shining brown. But for this, and the tooth-
knob, and the cavities of the keel, the whole scaphium is brilliantly white, reflecting
the sunlight like the albumen of a boiled egg. The lower portions, both of cheeks and
keel, run obliquely downward and abdomenward in strong longitudinal folds.

I know of no species in which the scaphium can be so effectively studied as this.
Yet I am compelled to confess an humbling uncertainty of its function. I carefully dis-
sected away the surrounding parts, and found this white organ to be firmly connected
with the chitinous base of the uncus (that is, with the eighth segment), and, apparently,
with a tube proceeding from the mass of convoluted yellow tubes which lie immediately
behind, which I suppose to be the biliary vessels of the colon. The orifice, however, is,
so far as I am able to judge, not at the extremity of this organ, as described by llerold
(if this indeed is his Tricmgelstuck, as I presume), but in the middle of the upper part of
its very basis, where it is united with the uncus. Projecting thus between the uncus and
the penis, we might call the scaphium a process from the perinaeum.

* But sec P. Demoleus, p. 313, infra.

f These great spoon-teeth are thus clearly identified with the aristate cheeks in 0. Remus, &c.

SECOND SEKIES. — ZOOLOGY, VOL. II. 43

306 ME. P. H. GOSSE ON THE CLASPING-OEGANS

Whatever its office, I infer it to be, in part at least, muscular, from its so often carry-
ing those remarkable weapons which I have called " double teeth." For, based on yield-
ing and non-contractile tissue, they would be ineffective and useless ; whereas their firm
chitinous substance, their depth of colour, and their high polish, indicating hardness,
and, as just described, their notched edges and acute points, all indicate vigorous and
definite action, dependent on strong muscular contraction.

Nevertheless, I am bound to confess, the position of these weapons and their direction
just under the uncus do not seem the most favourable for the only function which we
can attribute to them, viz. copulative prehension.

The penis is a straight, nearly horizontal column of brown clear chitinous substance,
moderately thick, not sensibly expanding, as usual, but rather diminishing to the extre-
mity, and ending in a linger-like poiut and irregularly lobed lips. I carefully removed
the walls of the abdominal segments, till I exposed the root of the cylindrical penis in a
mass of muscular tissue, just at the insertion of the sixth segment.

Papilio Bruits, Fabr. (Plate XXIX. figs. 2i: 25.)

The details that I have just been giving were from specimens labelled as having been
taken at Calabar, in West Africa ; and they agreed in markings, almost exactly, on both
surfaces, with the Merope of Cramer's plate 378. b, e. But I have examples of a style of
colouring that differs much from any that I have seen figured or described by authors.
These two males, ticketed P. Brutus, I purchased from a small collection made at Ellongo,
in Madagascar, arc distinguished by having the black spot at the anal angle of the hind
wing above very small, and the Avhole of the hind wing below, of a warm yellow-brown
hue, with an undefined light dash across the middle, and scarcely any trace of the dark
lines so conspicuous in the ordinary forms.

Now, in these Madagascar examples, the harpe takes a form very different from that in
the continental P. Merope. The valve has much in common ; and so has the harpe, in its
plan ; but the serrated portion, which is marked off by abrupt dilatation, is reduced to a
small ovate blade at the very termination ; and the projecting arm is placed close
beneath it, is swollen in the middle, sinuately curved, and tapered to an acute point.
The staff is very long and slender, cylindrical, pale, shining, bending from the base to
the dorsal corner, sunken in a deep abrupt depression of the valve-cavity. The head or
blade is free, a thin oval plate of polished black chitine, brought to an edge all round,
and cut into strong saw-teeth, which are divided by grooves that run up into the area
of the blade. Just at the point where it adheres to the lining-membrane, there is an
angular expansion of the staff which sends off, nearly at a right angle, the polished taper
arm, in quite a different plane from that of the blade ; this is not serrated.

This long sharp spine, and its fellow, stretch away on either side of the uncus, when
the valves are closed, and are doubtless effective prehensors ; whereas the short, blunt
arms, in Merope, placed so low down in the valves, one cannot well suppose to be of any
prehensile power. On the other hand, in Merope, the extent of serrated blade is greatly
superior.

The abdominal organs are the same, essentially, in both. For though, in one speci-

IN CERTAIN LEPIDOPTEEA. 307

*

men, the uncus was aborted, leaving only a slight irregularly pointed projection from
the tegumen, and in each the fleshy portions of the scaphium were partly gone, yet
in one case the uncus was perfect; and in each one pair of the " double teeth " was
left, and in one case, one of the other pair ; allowing to me to see that their form and
armature were exactly as I have described in Merope. In one example the entire soft
parts of the scaphium were lacking (perhaps eaten by Psoci), leaving the central mass
that supported the serrate teeth merely as a thin transparent shell of homy chitine.

I remarked that the abdominal cavity, in all the specimens that I examined, of both
the forms, and all the included organs, even to the fringe of the valves, were much
clogged with meconium and adhering body-scales, received, of course, from females.

Papilio Homeris, Fabr. (Plate XXX. figs. 1-1.)

The grand P. Homerus of Jamaica has a prehensile armature of much simplicity,
but quite unlike that of any other known to me. The valve is somewhat long, with
a rounded point bent-over towards the ventral side. It is dull, dark brown, clad with
small scales interspersed with hairs, externally ; internally it is simply hollowed, the scaies
scarcely protruding beyond the margin ; but a crowd of very fine hairs spring from
some distance within the margin, on both sides, and extend considerably beyond it.

The horpe consists of a simple rod of brown chitine, hollow throughout, very slender
at the base, widening abruptly in the middle into a llattened blade, which terminates
in two nearly straight points, of which the dorsal rises above the margin of the valve,
and the ventral dives into the hollow. The chitine is darker and, I presume, denser
in the vicinity of these points than elsewhere. The whole blade is sparsely beset with
long slender bristles. The harpe runs in nearly a straight line along the middle of
the valve, through its entire length ; but adheres only by the slender portion.

The uncus is long and stout, but has only a moderate downward curvature, very
regular and graceful, however ; the keel, produced by the united rami or vertical supports
of the sides, is deep. Viewed laterally, the uncus ends in a very sharp point ; viewed
posteriorly, it is a narrow, tapering spathula.

The scaphium is broadly boat-shaped, with little depth, and nothing that can dis-
tinctly be called a keel ; the " double tooth '; on each side is remarkably stout and strong
consisting of two conical cusps of shining brown chitine, united on a single basis, seated
on an ovate expanse, or dilated lip. These are angular-sided ; and the interior edge, espe-
cially of the hinder cusp, is beset with numerous short bristles, from which the extreme
points are free. A supernumerary spinous tooth, polished, slender, acute, erect, rises
from a knob situate centrally in the midst of the scaphial furrow, close to the extremity.

The p>enis is rather small and short, little protruded from the abdomen, chifcinous
brown, turgid at the extremity, hiaut, not expanded, with no finger-point.

Papilio Ulysses, Linn. (A. D.) (Plate XXX. figs. 5-8.)

Valve of more than ordinary dimensions; outline along semioval, bluut-pointed,
with a considerable sinuation on the dorsal side of the point. The black, coarse hair-
scales of the exterior project notably along the dorsal margin, and around the tip,

43*

308 ME. P. H. GOSSE ON THE CLASPING-OEGAJNTS

where they are longest ; then suddenly cease, giving place to a series of very fine shinin
hairs, which spring from the cavity, within the ventral margin, and reach heyond it.
The whole valve-floor, which is of a bright brown hue and has a velvety texture, is more
or less studded with gilded hairs.

The harpe is a slender rod, black and polished throughout, affixed to the lining-
membrane, parallel with the ventral margin nearly to the valve-tip ; then it makes
a sudden bend, and projects as a curved acute spine-like point as far as the ends of the
dorsal margin-scales. From the bend onward the rod becomes free, springing from
the cavity-floor, but proceeding laterally, nearly in its plane, the taper point describing
a full semicircle, curving at last quite upward. It is a simple polished wire, densely
set along its inner side with long shining hairs, but quite void of serration in every
part.

The abdominal organs present some peculiarities of form. The -miens is unusually
broad and short, taking the shape of a very wide but pointed spoon-bowl, clothed
almost to its termination with coarse, brown body-scales, among which individual blue
scales gleam Hke tiny sapphires.

The scaphiani is of moderate size, its most conspicuous feature being the lateral
teeth. The ridge, or cheek, on each side carries a stout and strong spine of polished
black chitine, which projects horizontally forward at first, and then bends up at the
point, the secondary tooth being represented by a low conical knob at the base. The
whole tooth is beset with long bristles, especially about the middle ; but immediately in
the vicinity of the point these are very few and minute. High magnifying power shows
(fig. 8) that, as in P. Merope, the scaphial teeth arc themselves toothed ; but here the
notching is shallow, and the minute bristles, instead of being seated on the summits of
the toothlets, are sunk in the insterstitial angles.

The penis is moderately large and protrusile, bending upward and outward its column
of brown translucent chitine, and terminating in a dilated, but not expanding, orifice.

Papilio Phokcas, Cram. (Plate XXX. fig. 9.)

Valve small, but comparatively long and narrow, and slightly fringed.
Harpe of the very simplest form, a mere narrow rod-like elevation of the chitine,
running, in a slightly bent line, throughout the length of the valve, and terminating in
a fine smooth point, Avhich protrudes even a little beyond the fringing hair-scales. The
whole rod is of polished dark brown chitine. (PI. XXX. fig. 9.)

Uncus well developed, long, slender, curved into a nearly complete semicircle. The
adjacent parts present nothing calling for special notice.

Here is a very close ally of P. Mreus, a West-African species (the specimen labelled
as captured at Accra), yet widely differing from both it and its nearer compatriot
P. Bromius, in the form of the genital armature. It approaches the Oriental P. Ulysses
and P. Arcturus in the form of the harpe, but exceeds both in simplicity.

Papilio Thoas, Linn. (Plate XXX. figs. 10, 11.)

In this noble American butterfly, so abundantly found from Virginia to La Plata, we
have a new type of valval armature. The outline of the valve is rotundo-triangular,

IN CEETAIN LEPIDOPTEEA. 309

moderately regular, the ventral side the longer, the whole margin broadly and pretty
evenly fringed.

The harpe consists of a very thin crystalline plate of chitine, which sits close upon the
central cavity of the valve. It is thickened on each side into two slender ridges, black,
polished, wire-like, which, springing from a wide basis, proceed in curving bines, and meet
in an acute point a little on the dorsal side of the extremity of the valve. The extreme
point separates from the lining-membrane, and projects at a very acute angle, firm and
hard, like the point of a needle. Under a high power the ventral side of the point is
seen to be notched into half a dozen minute saw-teeth : the whole triangular hollow plate
may be, with care, lifted uninjured from the valve ; it is then of a glassy transparency
and polish, slightly tinged with brown, like glass very slightly smoked ; very slender
veins issue from various points of the deep-black thickened marginal walls, and ramify
over the clear surface.

The abdominal armature of the first specimen I examined, seemed to be peculiar.
A deep, but short, uncus was seen descending perpendicularly and abruptly from the
tegumen, with almost no curvature at all, the upper surface hollow, from the turning-
up of the lateral edges, the tip slightly spathulate.

The scaphium was with difficulty recognizable, wide, shallow, and so short as scarcely
to reach the vertical uncus.

But a second example proved that this condition of the organs was a case of accidental
distortion, not very rare. The uncus now was normal in length and curvature, wide,
with a small spathulate point, followed almost immediately, in full depth, by a deep and
strong radial keel. (PL XXX. fig. 11.)

The scaphium is on the same pattern as in P. Mayo, still more closely like that of
P. Heleuus, both the lateral teeth being very stout, long, acute, dark, and polished. The
hinder tooth is rather a high conical incurved point on the thin shelly edge of the cheek,
whose summit bears a crest of small erect aristae. The edges are simple, not serrate.
Viewed from above, the cheeks take a subglobular, inflated form, and both the teeth of
the right side slope-over towards the median line of the uncus. The keel of the scaphium
is large but shallow, polished ivory-white, strongly defined in colour from the cheeks.
Its form is very boat-like, projecting its point horizontally ; the extreme point is minutely
complicate in surface, and diverse in colour and appearance from the contiguous parts.

The penis presents little noteworthy ; it is of moderate dimensions, a tube apparently
subcylindrical, but slightly enlarging to the tip, which ends in a lengthened curved
acute finger-point.

Papilio Turnus, Linn. (Plate XXX. figs. 12-15.)

Valve very regular in form, of the outline of a Gothic arch, the dorsal margin con-
cealed under long shaggy black hair, springing from within, and projecting far beyond,
the edge ; the ventral margin is also hairy, but less conspicuously. A flat shelf runs all
round, broadest at the extremity, within which the cavity abruptly deepens.

Harpe of the Thoas pattern, consisting apparently of a dark shining wire, running
up on each side, and meeting in an arch towards the tip. In reality, these are but the
thickened edges of a continuous plate of chitine, of extreme delicacy and glassy trans-

310 MR. P. II. GOSSE OX THE CLASPING-OBGANS

parency, which occupies the central area of the valve. Each wiry edge is armed with a
prehensor ; the ventral edge, a little beyond its middle, bears a long acute spine, which
stands up from it, and curves obliquely forwards ; the dorsal, a thicker cone, more erect
on its base, which first sends off a nearly straight horizontal spine obliquely forwards, and
then terminates in a shorter and blunter point.

The uncus arches in a regular semicircular curve to a small spathulate point ; a deep
keel beneath the receding rami ; between which issues the moderately large scaphium,
well keeled, the only peculiarity noticed being that the extremity of the keel, by its
deepening hue, appears to become more chitinous, until it ends in a minute horny
point. The double teeth are minute, but distinct, conical points.

The penis is small and high-situate.

Papilio Polvcaox, Cram. (Plate XXX. figs. 10-19.)

T'a Ice-outline subrhomboidal, with rounded angles, thickly fringed on the dorsal
(short) side, but thinly on other parts.

Harpe a broad, concave plate of thin, clear, shining chitine, beset with fine, short
bristles, pointing outward, as is also the lining-membrane where it is exposed. The ven-
tral edge of this plate is a slender nearly straight rod, polished, black, with a fine-drawn
acute point, free for some distance. The upper edge of the thin plate usually springs
from the spine a short distance down, and rises into a wide curve ; sometimes it follows the
spine up to its point, and descends at an acute angle ; but presently, on what I may call its
dorsal corner, it gives rise to a curious appendage, which usually takes the form of a thick,
short, solid column of black chitine, set, all round its summit, with stiff tooth-like spines
(fig. 18) ; but sometimes appears as a smaller knob, with a slender base, whence rise three
diverging lamina?, united by their inner, and serrate on their outer edges (fig. 17). In
either form, the spines are of a clear, glassy, blue-black substance, and most of them are
double, *. c. each has two points from the same base. The spinous rod of the ventral edge
bears a triangular thin elevation (and sometimes a secondary one) near the base.

The uncus is long, slender, evenly curved, acute; the keel of its united vertical rami
deep, and diminishing in depth gradually to the tip.

The scaphium is very ample in all its constituent parts ; the cheeks rise into high
arching crests, with notched, aristate edges, but destitute of double teeth, so far as I can
discern ; the sides run down on either hand in great oblique folds ; the keel is pointed,
large, and deep; all the parts are shining white. The penis presented no features
worthy of special notice, other than can be observed in the figures.

In my cabinet are examples of six very diverse kinds of butterflies, all large, imposing,
and richly coloured, labelled thus : — Androgens, Cram., 6 and 2 , from Brazil ; Thersiles,
Pabr., 6 and 2 , from Jamaica ; Polycaon, Cram., 6 , and Lyeophron, Hiibn., 2 , from
Corrientes, in the Argentine Republic *.

* Those may be readily distinguished thus: — Males: 1, hind wings with no yellow hmules= Androgeos ; 2, with
shallow yellow kmvles=Thersites; 3, with deep yellow lumiles= Polycaon. Females: 1, fore wings with a broad
straight yellow bar across =Andror/eos : 2, with a narrow curved yellow band across =Thersites; 3, no yellow band
across = Lyeophron, i.e. Polycaon $ . I describe from my own specimens, which differ from Cramer's figures. But
what I call Polycaon may possibly be the Oebalus of Boisduval (Sp. Gen. p. 3G0), with whose description they fairly
agree, except that I see no trace of the " red-brown lunules " that he assigns to the hind wings above.

IN CERTAIN LEPIDOPTERA. 311

Xow the genital armature, in these widely severed forms, is essentially the same.
Of that from Jamaica, and that from Corricntes, I have received a considerable number
of examples, taken at different times by my own friends ; so that I am sure of the habitats.
The Jamaican specimens have the isolate knob of the harpe three-sided or concave,
the main spine slightly curved, and with but one baseward elevation. The Androgeos
from Brazil, and several Polycaons from Corrientes, show absolutely no difference inter se.
The characters are constant, so far as my experience goes.

The precision with which this very peculiar harpe is repeated in these widely spread
varieties is something noteworthy. It suggests the notion that, in this organ, we have
a specific character of much value. Androgeos, Thersites, Polyraon,xmA. Lycophron, are
assuredly but a single species.

Papilio Axiox, Boisd. (A. D.) (Plate XXX. fig. 20.)

The armature of this noble butterfly is remarkable. The voice is almost as large as
that of an Ornithoptera, and of similar shape, viz. semioval, or of the outline of a half
lemon cut obliquely, the dorsal side by far the shorter, as usual. Exteriorly, the valve
is clad with dull white scales, which become black at the dorsal edge, and these pro-
ject pretty evenly beyond the margins. Interiorly, it is umber-brown, with a dull
gloss.

Prom one specimen, on removing a valve for examiuation, the whole cavity was
fouud quite full of a light brown dust, composed of amorphous fragments (of meconium,
surely), together with a considerable number of fine hair-like scales, such as are proper
to the hinder regions, gathered, I doubt not, from the body of the female during an act
of coition.

The harpe belongs to the pattern of Thous and Turnus. It is an ample plate of thin
glassy chitine, which closely lines the greater part of the entire cavity ; yet can be easily
separated from it, in its integrity, when it is a curious and attractive object. It is a
broad hollow plate, of exquisite delicacy and tenuity, translucent, slightly tinged with
horn-yellow, the extreme edge, all round, thickened and deepened in colour, polished,
■ and furnished at two points with two long, taper, curved, acute spines, much resembling
the prickles of a rose in form, colour, and polish. Both are on the ventral margin, the
one just below the apex of the valve being the proper prehensor, laterally opposable,
conjoined with its fellow in the other valve, to the uncus. Besides these spines, there
are several minuter ones on the dorsal edge, the chitine rising here into a thin wall,
which is cut into teeth, three or four, the number and arrangement being not quite
the same in two individuals that I examined. The form of the whole harpe varies also ;
for while, in the specimen figured, which was the first I examined, the apical arch was
somewhat flattened (as represented), in a subsequent example the outline was much more
graceful, more nearly corresponding to that of the valve, and so more closely resembled
the form in Thoas and Turnus.

The penis appears to be small, short, and curved into a semicircle ; but it was so closely
enveloped in its ligaments, that I could not very satisfactorily examine it.

The uncus with its concomitant parts presented nothing specially notable.

312 ME. P. H. GOSSE ON THE CLASPING-OEGANS

Papilio Hesperus, Westw. (Plate XXX. figs. 21-24.)

Valve of the outline of a Gothic arch obliquely cut off at base, wholly surrounded by
a dense and wide fringe of grey-brown hair-scales, densest and widest at and around the
tip. Interior surrounded by a wide flat shelf, within which the cavity abruptly descends,
and is deep.

The harpe occupies the whole of this deep area, and fits it accurately with a chitinous
hollow plate of extreme delicacy and transparency, thickening and darkening all round
into a narrow edge, which stands up free, like a little wall, the summit rising into rounded
eminences here and there (PL XXX. fig. 21). The structure could not be detached in
integrity ; the ventral portion of the circuit is represented at fig. 22, viewed from the
ventral side. It is now seen that the free edge is cut into saw-teeth, and these by no
means simple, but elaborately compound, each tooth, at least for a considerable portion
of the edge, being itself cut into three, or even four toothlets, and each of these com-
pound teeth being the termination of a rounded curved ridge on the outer surface of the
wall. I have found no structure elsewhere quite analogous to this. Fig. 23 represents
some of the teeth on a more enlarged scale.

The abdominal apparatus is of the ordinary type ; the uncus is moderately long, in-
clining downwards, nearly straight, crowned with a ridge of stiff erect hairs, deeply
keeled. Scaphium moderate, double teeth strongly developed, erect, incurved, acute.
Penis doubly curved, finger-pointed, moderately thick, and so long as to project hori-
zontally between the valves when these are closed.

Papixio Demoleus, Linn. (Plate XXX. figs. 25-27.)

Valve-outline an oblique semiovoid ; exteriorly clad with a coat of drab-white, short
scales, which project in a very slight fringe just around the extremity. Interiorly the
surface smooth, but not shining ; the cavity rather narrow, running to a point, bounded
by a broad flat shelf on each side, of which the dorsal extends only half down, the ventral
the whole length. ;

Harpe a broad thin plate of yellow transparent chitine, not very concave, reaching to
within two thirds of the valve's length ; its ventral edge an elevated ridge, apparently
tubular, with an expanded base ; the ridge, like a clear glass tube, is marked, at short
intervals, by ten or twelve distinct black transverse lines, suggesting joints, or rings of
metal binding a pipe for increased strength. Close to, but not quite at, the extremity
of this ridge, which diminishes to a thin scalpel-like edge, there is a single minute black
tooth projecting obliquely towards the dorsal side, the upper edge of which, in the
specimen examined, showed a few minute notches, possibly accidental. The area of the
harpe is corrugated into fine ramifying wrinkles, and studded with very minute circular
areola? (pits ?), from each of which springs a hair, varying in fineness. The lining-mem-
brane of the valve is similarly studded, the areolae and hairs specially crowded toward
the point. The fringing hairs (hair-scales) are in the same plane as the valve-surface ;
these perpendicular to it. Besides these hairs, there lie, about the area of the harpe,
many body-scales, of varying forms, which, derived doubtless from the other sex, usefully
illustrate the function of this organ.

IN CERTAIN LEPIDOPTERA. 313

The abdominal apparatus is normal, and well developed. The uncus is moderately
long, well curved, slender to the tegumen, terminating in a minute spathula ; the keel,
reaching to the point, is moderately deep, the edges not upturned.

Scaphium of the pattern seen in P. Merope. Its keel is more produced, and of less
irregular surface. Its " double teeth " are not quite so much developed, particularly the
secondary one of each pair, which is little more than a conic knob, both, however, set
with prominences and bristles as in that species, though in a less degree.

Penis of moderate length and thickness, increasing to the extremity, which is truncate,
hiant, without lateral expansions.

Papilio Batuycles, Zink. (Plate XXXI. figs. 1-5.)

Here are very abnormal structures.

The valve is small, short, almost semicircular, widely fringed with dense white hair-
scales. The harpe is a broad segment of a sphere, of glassy chitine, thickened at the two
margins, and rising at the summit into two divaricating branches, each of which, the
dorsal especially, is studded with short stiff blunt teeth ; on the dorsal edge, near the
base, there is also a conical process. Parts of the surface are beset with fine scattered
bristles.

There is no uncus proper. The regions are sheltered by a dense horizontal thatch of
long white hair-scales. "When these are removed (by the persevering friction of a camel's-
hair pencil cut short), we get an instructive sight of the organs. The median line of
the eighth segment projects into a very shallow point, from under which a well-formed
white scaphium is moderately developed. At fig. 3 is shown the transverse terminal
line of the eighth segment, and a vertical view of the scaphium ; the lip-like lateral
edges bear no trace of the ordinary " double teeth." At fig. 4 is given a lateral view of
the organs. The vertical dotted line shows where the lateral edge of the dorsal arch of
the eighth segment falls normally ; but this has been removed, so that we can still
follow the outline of the scaphium somewhat further into the abdomen. The median line
of the arch is supposed to be retained, forming the roof of the cavity ; and to this we
perceive the scaphium is organically attached, descending from it in a great angular
heel, in the free interior, and projecting, in the familiar form, from beneath and within
the spreading rami of the uncus, where this is (as is usually the case) present. From the
median part of its inferior surface there descend membranous lamina! of irregular shape,
perforated to allow the exit of the penis. Here this organ, seen in situ at fig. 4, and its
tip, more magnified, at fig. 5, is of great length. It appears to be invested with a thin
glistening membrane, but only partially throughout its length, expanding and dilating
near the tip, which, furnished with a minute finger-like point, is free. In the normal
condition it protrudes far beyond the limits of the closed valves.

Thus every one of the organs is remarkable. The valve, the harpe, the uncus (absent),
the scaphium, the penis, all present something unusual. The very simplicity of the
scaphium, its lack of teeth, spines, or bristles, seems to deny the " reason of its being."

SECOND SERIES. — ZOOLOGY, VOL. II. 41

311 MR. P. II. GOSSB ON THE CLASPING-OEGANS

Papilio Agamemnon, Linn. (Plate XXXI. figs. 0-8.)

Valve small, sliort, rotundo-triangular, fringed with long hairs, not very dense. There
is a broad, fiat shelf, going nearly all round, except at the base; this shelf is beset with
stout but short hairs, which bend over the inner edge; and the very edge itself is fringed
with short stiff bristles, almost like spines.

At the base of the cavity lies the harpe, a shallow cup of chitine, thickened around
the edge, and slightly free ; on its dorsal side it rises into a tooth-like lobe of thin but
dark brown chitine, of rotundo-triangular outline, which docs not appear to be serrated
(PI. XXXI. fig. G).

The circumference of the abdominal cavity is densely set with projecting hair-scales,
which in particular rise from the summit of the ultimate segment to its very point, and
then, arching forward, form a long horizontal thatch. When these hairs are detached
(as. in part, at fig. 7), the point, ordinarily the uncus, is seen to be short and slightly bent
upward, and, viewed vertically fas at tig. 8), to be truncate, and even very slightly bifid
at the tip. Below, there is a well-formed but very small scaphium, with no armature;
and then a penis of much slenderness, and so long thai its point protrudes from between
the valves, when these are closed. The chitinous portion is downward, ends in a point of
greal tenuity, and forms an imperfeel tube with arching edges above, within which,
prominent near the tip, is a pulpy white tissue. The sheath of laminae descending from
the scaphium is conspicuous and shaped like an elegantly pointed leaf on each side,
united below.

Papilio Erithonius, Cram. (Plate XXXI. figs. 9-12.)

Valve of an outline somewhat trapezoidal, with the angles rounded, the terminal
angle produced, sometimes truncate, and ending in three small obtuse lobes. A fringe
of hair-scales beyond the margin on the dorsal side only; the floor of the cavity,
where it is not covered by the harpe, having a cloth-like surface, unusual.

Nearly allied, as this Indian butterfly seems, in form and colouring of the wings, to the
African P. Demoleus, there is no resemblance in the appearance of its valve (see PI. XXX.
fig. 25). Nor is there any more in its harpe ; for this organ is a shallow cup of chitine,
quite rilling and accurately fitting the basal half (and more) of the valve-cavity, having its
ventral edge thickened, and the dorsal rising into a thin chitinous wall, of considerable
height, the summit of which is serrated with sharp teeth (PL XXXI. fig. 9). When
this harpe is carefully separated from the lining-membrane of the valve, it is seen to
be less simple ; for the elevated dorsal lobe is now seen to consist of three parallel
lamina?, which are of various lengths and heights, of which the inmost (also the tallest)
alone is serrated.

In the abdominal apparatus there is a general agreement with the Merope and
Demoleus pattern, with considerable diversity in detail. The uncus is short, little curved,
ending in a broad and thick spathulate point, the keel deep, the rami rapidly receding,
rising above the level of the summit, and forming a wide triangular arched roof to the
cavity. The scaphium is minute, but possesses the essential features ; the " double tooth "
on each side is well marked ; the frontal keel, however, is replaced by two erect fleshy

IN CERTAIN LEPIDOPTERA. 315

points. The penis is somewhat thick, moderately long, blunt-pointed, turgid, the white
pulpy tissue copious.

At fig. 11 these organs are represented as viewed sidewise, the surfaces having been
denuded of the copious scales and hair-scales, and both the valves removed. At fig. 12 is
shown the corneous skeleton of these regions. By one of those fortunate accidents uf
which an observant scient is always so glad to avail himself (as he can never command
them) a specimen of P. Erithonius, which had been badly infested with mould, came to
pieces in my hand, and the terminal segment dropped away from all the debris, clean,
just as I have sketched it at fig. 11. Here I saw not only the uncus terminating the final
segment, and the scaphium attached to it beneath, but, lower down, the orifice for
the extrusion of the penis, pierced through the middle of a vertical curtain of very
thin horny chitiue, which is suspended in place, by attachments on each side, to the
lower parts of the dorsal arch of the eighth (?) segment, as well as by two perpendicular
ligaments below, which tie it fast, the one to the middle of the ventral arch of the ninth,
the other (wider and thinner) to that of the eighth segment.

Perhaps — since I have never met with any thing like this structure, before or since —
it may not be impertinent to add that I have not exercised my imagination upon a
piece of integument irregularly pierced and fractured ; the screens, the orifices, and their
edges were quite symmetrical and smooth, exactly as I have represented them ; and this
chitinous curtain is still in my possession, in excellent preservation.

Papilio Ancuisiades, Esp. (Plate XXXI. figs. 13-15.)

Valve nearly parallel-sided, bluntly arched at the tip ; margins furnished, more or less
completely, with a narrow fringe of hair-scales. The cavity is abruptly depressed a little
way within the margin on each side. Just at the very edge of the descent, on the dorsal
side, I observed in one specimen two minute black teeth curving towards the cavity ; but
of these I could find no trace in other specimens.

The harjpe is a thin concave plate, long, narrow, parallel-sided, seated within the
depression, but not reaching either of its margins. It can be easily lifted unbroken and
laid on a slip of glass ; and is then seen to have a continuous floor of yellow chitiue,
excessively thin, and yet apparently of two surfaces, since the unequal contraction
of drying has raised irregularly ramifying, thickened, fine lines, meandering like rivers on
a map ; Avhile the bounding lines, and the upper part of the arch produced by their
meeting, are of a deep, rich, yellow-brown hue, gleaming and glittering in the changing
light, like cut glass. The abrupt termination of this darker part of the arch, within,
conveys the idea of an overhanging edge : but this is delusive ; for, by careful mani-
pulation under the microscope, I proved that this etfect is produced merely by increase
of density. The point of the arch is a long taper tooth ; and three or four teeth on each
side descend along the outer edges of the arch. But the minuter details of number and
order are not constant in different examples. The dark tooth-like objects which I have
represented within the point, directed inward, are not true teeth projecting from the
chitinous surface ; for, though they appeared exactly as I have drawn, yet when I slowly
and carefully tilted the object under the microscope, so as to obtain changing angles of

316 ME. P. H. GOSSE ON THE CLASPLNG-OEGANS

vision, there was no break of the light reflected from the shining surface. A.nd in other
examples there was no trace of them. These appearances, therefore, I cannot satisfac-
torily explain.

The varus is curved in the usual arc ; it is thatched, for most of its length, with
coarse shaggy black hair ; the extremity forms a thick rounded spoon, bent abruptly
downward ; the keel produced by the vertical rami is unusually deep.

The scaphivm is boat-shaped, rather shallow, with a cleft tip, and no proper keel, tbe
sides are full and round ; their summits, brown with longitudinal chitinous thickenings,
have no trace of the normal " double teeth," but are surmounted with ridges of long stiff
bristles.

The penis is of moderate dimensions and protrusion, without terminal expansion.

Papilio CniLURKX.1% Cray. (Plate XXXI. figs. 10-18.)

Valve sabtrigonal, pointed, nearly equal-sided, sparsely fringed with hair-scales, and besei
with hairs, pointing inward from the ventral margin, and specially crowded near the tip.

Harpe nearly coextensive with the valve, springing from the entire base, and forming
a concave triangular plate, with both lateral edges thickened, and a strong ridge running
up near the middle. The ventral edge rises from an expanded hollow base, like the
trunk of a tree, sending oif near its middle point a stout sharp erect spine ; at its end,
where the central ridge meets it. both expand into a sort of hollow hand, turned up and
curved-over ventrally, having the short fingers of glittering black chitine on half of its
semicircular edge, the rest being a smooth wall. The dorsal edge of the harpe stands
close to the edge of the valve, from which it rises into a thin wall of considerable height,
leaning inwards, fringed along its summit, bearing a row of fine hairs.

The chief feature remarkable in the abdominal organs is the penis, which is rather thick,
and so long as to protrude beyond the closed valves ; its hue is pale chitine-yellow ; its
expanded orifice is filled with the white pulpy tissue, which swells beyond the margins;
and there are, on the column, successive marks, which are repetitions of the terminal
outline. (See the same organs in Orn. Raliphron, p. 285, supra; PL XXVI. fig. 7.)

Papilio Sesostris, Cram. (Plate XXXI. fig. 19.)

The general type is that of P. Childrena; and the similarity is greater (considering
the complexity of the armature) than I have observed in any two forms which may
be considered species, or even marked varieties. Yet there are numerous diversities
between these two.

The valve is of a different outline, parallel-sided and rounded, instead of trigonal
and pointed ; it is more distinctly fringed, along the dorsal margin and around the
end, with close-set hair-scales.

The harpe, while of the same pattern, is much modified. Its ventral edge is straight,
the solitary spine placed nearer the base, and reduced to a wart. The ridge is now the
principal feature, which in P. Children^ is secondary ; it is more curved. The hand-like
tip is similar in form ; but the short blunt fingers, curving over the cavity, are replaced
by sharp teeth, curved like the spines of a rose, nearly in the plane of the valve (as

IN CERTAIN LEPIDOPTERA. 317

shown in outline in PL XXXI. fig. 19). The wall-like dorsal plate does not end
ibruptly, but is continued all round the extremity of the valve, and even far down the
rentral margin, so that the harpe seems to be quite coextensive with the valve. Theu,
)etween this wall and the main ridge, there is a secondary ridge, of which just a
;race appears in P. Childrence, but which is here conspicuous ; membranous for the most
)f its length, but supported, near the part where it curves up behind the " hand," by a
eaniug " strut," of deep-brown shining chitine.

These are differences which would require representation on a much increased scale,
o make them very appreciable by figures, but which are striking enough on careful
'xamination.

When I came to examine the abdominal apparatus in P. Sesostris, I at once found a
very glaring aberrance from P. Childrence. Por there was not a trace of an uncus, the
edge of the upper arch of the eighth segment projecting in only a very gentle curve,
beneath which the arch of the ninth was wholly out of sight ; the mingled long bristles,
black and crimson, that generally project as a tuft, horizontally between the valves,
were here bent down vertically into the cavity ; and, behind them, a very minute hook
was projected. There was, moreover, but the slightest doubtful trace of a scaphium, and
only when I threw the sunlight far into the abdomen. The penis, in size, length and
direction, agreed fairly with that of Childrence, but no white pulp was visible; the
whole organ was of an uniform deep shining brown ; and instead of expanding, its
thickness gradually diminished to the tip.

All the diversities which I have enumerated, minute individually but in cumulation
great, seem to me to point to original specific distinction, though the evidence is no
incontrovertible.

Papilio Vertfmxus, Cram. (Plate XXXI. figs. 20, 21.)

Vali-cand harpe agreeing in plan with those of P. Muryleon (infra), and more essentially
with those of P. Sesostris and P. Childrence, particularly the latter. The order of
resemblance is — Childrence, Sesostris, Vertumnus, Euryleon. The two ridges, enclosing
a triangular area, are gone in the latter two, and the harpe is a chitinous plate of about
equal width throughout. In all, there is the high and broad tooth in the middle of the
ventral side, followed, in the latter two, by smaller teeth ; the terminal hollow " hand "
of P. Childrence is best represented in P. Vertumnus, where the form is similar but wider,
and the fringing overcurving teeth are eight.

The abdominal organs are normal. A long slender well-curved uncus ; an ordinary
deep-keeled seaphiuin, with horizontal "double teeth;" and a moderately large penis,
with expanded orifice, seated far back.

Papilio Lysaxder, Cram. (Plate XXXI. figs. 22-24.)

Valve semiovoid ; very slightly fringed ; a flat shelf within the ventral margin.

Harpe a broad triangular plate, that occupies a great part of the breadth of the valve.
Prom an ample expanded base, by which it was attached to the supporting knob, it springs,
tree-trunk-fashion, with wide and thick ridges of glittering chitine, to the summit of the

318 ME. P. H. GOSSE ON THE CLASPING-OEGANS

valve, narrowing as it proceeds. Here another wide branch is united to it, or it may he
considered as a descending piece, since it has no affixed hase ; it skirts the dorsal margin.
The nnion of the two pieces seems perfect, though the substance of the secondary piece
becomes exceedingly attenuate there. Abroad space, well defined, occupying the middle
of the valve, is likewise so exceedingly attenuate, that it is only when we lift the harpe
out of the valve, that we perceive, by the continuity of the plate, that this median region
is chitinous at all. The extremity of this compound harpe runs to a sharp point, and
is notched into four teeth, of which the two on the dorsal side are double. But the
arrangement and form of the teeth are not quite constant.

The uncus is slender, abruptly bent near the base, thence straight but oblique, to the
terminal hook. A small scaphiwm, with difficulty recognizable, was (in one example)
attached, very closely appressed to its inferior surface, through its length ; the proper
" double tooth " on each side was represented by a close-lying horizontal spine, dark,
polished. But, in two other examples, the scaphium was wholly absent, or apparent only
as a minute amorphous rag of white flesh ; while, from the middle of the uncus, descended
vertically a divaricating fork of two spines, organically (as it seemed) jointed to the two
sides of the uncus. In each case these depending spines were the outer margins of veils
of pellucid white flesh ; and in one case they inclosed white matter, which appeared
identical with the pulpy substance of which I have already spoken. For here, again,
the penis, which was much distended, had its interior filled, to overflowing, with the
shining white substance, which not only occupied its whole gaping orifice, but was carried,
as a great ball, upon its extremity.

In another example the penis was very large (see fig. 21), diminishing to the extremity,
which was furnished with a hooked, slender finger-point ; the outer chitinous integument
was split through all the length of the organ that was visible ; and (what seemed note-
Avorthy) laterally and asymmetrically; so that a wide, gaping slit, obliquely, along the
side, looked as if the integument had been violently burst by distension within. As if
to confirm this conclusion, the tube appeared perfectly empty ; my eye, with the aid of
sunlight, could penetrate through the hiant cleft, far up into the interior, and all round ;
but could not detect a trace of what sometimes is so prominent, the second tissue ; no-
thing at all, but the thin, transparent, gall-yellow, tubular wall of chitine.

We have surely much to learn yet about these obscure organs, and their functions !

Papilio Euryleox, Hewits. (A. D.) (Plate XXXI. figs. 25, 20.)
Valve nearly semiovate, narrowly fringed ; the cavity bounded by a marginal shelf,
which runs up the ventral side, and a little way round the extremity.

The harpe has much resemblance to that of P. Anchisiades. It is a rather narrow,
parallel-sided, thin, concave plate, both sides of which are bent round dorsally ; and the
sharp tooth-like points, which in my examples were seven in number, are almost all on
the ventral edge, extending from the tip about halfway down. All are overcurvcd, hard,
and horny ; the lowest by far the largest. In removing the valve from its attachment,
the expanded tissue by which the harpe had been affixed to the basal knob, occasionally
so conspicuous at the bottom of the abdominal cavity, came away uninjured ; and when

IN CERTAIN LEPLDOPTERA. 319

I lifted the harpe from its valve with a scalpel, this basal expansion still came away,
as represented in my figures. It appeared shrivelled, thin, and membranous ; but a short
maceration with water caused this tissue to swell and become plump and soft. It is, I pre-
sume, a muscular attachment for the special energetic movement of the prehensile harpe.
Maceration produced not the slightest sensible change in the polished chitine, either now
or any other occasion.

Papilio Ergeteles, Gray. (Plate XXXI. figs. 27, 28.)

TVe have here another modification of the P. Anchisiades pattern of harpe. The ventral
side runs up through the very central line of the valve, having a strong tooth projecting
from its mid-wall, and three teeth at its extremity, whence a slender branch descends
ventrally, as in P. Lyscmder. Then, also, on this edge of the valve, as in the same species,
accessory pieces stand up as two half-walls, one within the other, of dark chitine.

The uncus and scaphium are both normal ; the latter specially well formed, though
small, with " double teeth." The penis, large and uncouth, much as in Lyscmder.

Papilio Macedon, Wall. (Plate XXXII. figs. 1-3.)

Valve of similar outline to those of P. Ulysses and P. Arcturus, yet sufficiently diverse
in other respects. It is scarcely fringed, except very thinly along the ventral edge ; but
the whole cavity-floor is studded with minute pits, from each of which springs a fiue
short hair, some of which exceed the margin.

The harpe is quite peculiar ; at the first sight it reminds us of the form common in
the Ornithopterce, the arm and the bent, spine-studded hand ; but, when detached, and
examined separately, as shown at PL XXXII. fig. 2, the armature is seen to be different,
and even yet more formidable. If I may repeat the comparison to a hand, the five
fingers, of glittering black chitine, are bent over the palm ; and each of the five is
notched, all down its front, with sharp teeth, of which the uppermost are the longest
and stoutest. This and its fellow must make a very effective pair of graspers !

The abdominal organs are all well developed. The uncus arches nearly to a semicircle,
terminating in a delicate spathula. The scaphium is definitely separated into two portions,
the upper, altogether of a rich brown hue, apparently chitinous, which bears the teeth,
and the lower, uniformly pellucid white. This latter is the keel, though it is not at all
produced. The former, in the place of the usual " double tooth " on each side, has but
one, erect, acute, recurved, like a viper's fang ; but the termination of the scaphium is cut
into three abnormal* teeth.

TXxa penis is very long, and somewhat slender, ending in a produced finger-like point, and
not much expanded. It bends obliquely downward, protruding so far as to reach, if it
does not exceed, the limits of the valves.

The harpe was, when I opened the valve, much choked with scales, doubtless from
some female.

Another example agreed very closely with the above, save that the teeth of the scaphium
were nearer the tip ; the inner edges were set with low points, each surmounted by a

320 ME. P. H. GOSSE ON THE CLASPING-OEGANS

bristle ; and the secondary tooth was represented by a rounded eminence, on which there
was a crest of stout glittering arista-.

Papilio Bromius, Doub. (Plate XXXII. figs. 4-G.)

Valve almost circular, margined, at the ventral edge and at the extremity, by dense
fringes of coarse hair-scales, and, along the dorsal edge, by sparse fine long hairs.

Harpe, a bold triangular hollow plate of chitine with thick edges, springing from an
expanded base, which almost fills the basal portion of the valve-cavity, and then, rising
erect, throws forward a broad truncate sharp edge, cut into four great teeth.

Uncus short, triangular, horizontal, strongly and deeply keeled. Scaphium large and
conspicuous, the keel bent obliquely downward, the "double teeth" replaced by a long
and high crest on each side, having several points. Penis of excessive length, slender,
wire-like, bent downwards, and conspicuously protruding between the bottoms of the
valves, in ,siln ; extremity shaped like a horse-hoof.

This is one of the must instructive cases that I have met with. Having examined the
South-African P. Nireus in three examples, and found no difference in the structure, or
arrangement, or relative size, of the parts, in the three, I thought of the West- African
P. JBromius. The two are so much alike, in every apparent feature, as to warrant the con-
clusion that they are but "local variants " of one and the same species. But lo! here,
in these hidden organs of a function of the most vital importance to the creature, I find
with amazement a radical aud very extensive diversity.

In one* the valve is trigonal, in the other circular; in one the harpe is a long spoon,
studded with cone-points ; in the other a horse-head with four teeth ; in one the uncus is
long and of extreme tenuity, in the other short, deep-keeled ; in one the scaphium bears
a pair of blunt erect pegs, in the other a series of compressed ridges ; in one the penis is
short, elevated, and terminates in a finger-point, in the other very long, depending,
unarmed.

Can these be descended from a common parentage ? and are the diversities merely the
result of changes in the climate, soil, and food produced on a party of emigrants, in the
course of many generations ? Or are they not, rather, powerful, if unexpected, witnesses
to the primal diversity of Papilio Nireus and Papilio Bromius, as distinct creations of
the Almighty God ?

Papilio Podaliritts, Linn. (Plate XXXII. figs. 7-10.)

The valve is thin, narrow, sharp-pointed, more developed on the dorsal than on the
ventral side, the latter remarkably excavated toward the base; the fringe of hair-scales,
moderate at other parts of the margin, is so enlarged around the excavation that the
irregularity of outline is not readily noticed (see PI. XXXII. fig. 7).

A hnpe, having a base as broad as the whole base of the valve, narrows at about the
middle, where it terminates in a summit of dark chitine, shaped somewhat like an eagle's
head and bcak; looking towards the ventral side. Prom the thinness of the valve, and
from the readiness with which the harpe comes away, the latter is favourable for

* See p. 32S, infra.

IN CEETAIX LEPIDOPTEEA. 321

microscopic examination (see fig. 8), It is now perceived to be a broad concave plate,
rising into somewhat thicker Avails at the sides, and having an elevated thin ridge running1
up in the midst. The "eagle's head " is very sensibly thickened, and abruptly, so as to
assume the semblance of a twofold wall of the tissue ; but very careful manipulation
has convinced me that this appearance is illusory. All over the thinner region meander
those irregular lines which I have noticed before, so much like rivers on a map, and
which I suspect are caused by the shri veiling of the investing tissue in drying. Besides
these, there are a few scattered bristles, each rising from a minute elevated knob or bulb.
The very edge of the "eagle's head" is a little arched-over; its line is uneven, but
I cannot detect any serration in any part.

There is no trace of an uncus. The posterior margin of the upper arch of the eighth
segment, when viewed vertically (as at fig. 9), is nearly straight, or follows an uniform
slight curve, with no attempt at central projection. Immediately beneath its middle
is seen a small but distinct scaphium, with the "double tooth" strongly developed in its
principal member, black, long, nearly straight, horizontal, reaching to the extremity of
the scaphium, the secondary visible only as a low shining knob. The sides of the
scaphium are rounded, concave exteriorly ; and a very distinct thin white keel is below.

Dr. Buchanan White, in his beautifully illustrated memoir, has given a vertical figure
of what he considers the tegumen of P. Podalirius (pi. lvi. fig. 39 b), which, in his text
(p. 361), he says is " almost bilobed." But I venture to think he has fallen into error
here, careful lateral examination showing that this structure proceeds from beneath the
straight edge of the tegumen, as I indicate at fig. 10. It is certainly, therefore, the sca-
phium, as the whole details prove.

The penis consists of a blade of glittering black chitiue, folded longitudinally, and
ending in a slender finger-point, with a knob at its origin. Thus folded, the sheath does
not half contain the white pulpy portion that runs along apparently within it, and at the
tip expanded beyond the finger. Possibly in life the firm chitinous sheath is semi-
tubular, and takes this folded form by drying. It sheds some light on the curious appear-
ance of the organ in P. Codrus (infra).

I have examined minutely two examples of this species, and find no appreciable differ-
ence in any of the organs observed.

Papilio Latreillii, Don. (Plate XXXII. fig. 11.)

Valve regular in outline, ovate, nearly equal-sided, pointed ; the exterior clad with a
dense coat of close-set scarlet hair-scales, which project, in an even thick fringe, beyond
the ventral margin throughout, but not sensibly on the dorsal margin, of which a few
black hairs only occupy the terminal moiety. Interiorly it is deeply and evenly hol-
lowed ; along the midst of the cavity runs an irregular-shaped harpe, adherent to the
lining-membrane (though readily detached), dilating into a broad blade of brown
chitine, serrated with minute teeth along its edge, and projecting a prominent angle at
each end. This blade is free, and stands up at an oblique inclination from the cavity.

The intravalvular organs I have examined imperfectly, and only in situ. The uncus

SECOND SERIES. — ZOOLOGY, VOL. II. 45

322 -AIR. P. H. GOSSE ON THE CLASPING-ORGANS

appears long, slender, and circularly arched, much as in P. Demoleus ; the scaphium
large, truncate, unkeeled, somewhat as in P. Moments.

Papilio Semperi, Peld. (Plate XXXII. figs. 12-14.)

The valve is moderately large, of irregular outline, having prominent rounded angles,
and a sharp projecting tooth at the summit. Exteriorly it is clothed with a coat of short
scales, which are black from the base to the middle, and scarlet thence to the tip.
Interiorly, the margins are exceeded by a broad fringe of hair-scales, which help to fill
up, and so conceal, the irregularities of the outline.

The harpe is a shining chit inous plate, running through the length of the valve mesially,
projecting two broad and sharp angular points at the extremity, and one in the middle of
the ventral side. PL XXXI I. fig. 12 shows the valve with the harpe in situ ; but if the
harpe be removed, which is done with great ease and perfectness, and viewed almost
along the plane in which it lay in the valve, its form is very different; for the terminal
portion curves upward, so that the points stand upright, and enclose a great semicircular
area, as shown at fig. 13.

The uncus presents little that is notable : it is nearly straight, with the point sharp
and bent down, bill-hook fashion; and its scaphium is rather small, and destitute of
" double teeth." The penis is of extraordinary dimensions, the triangular dilatation of its
extremity being enormously developed. At fig. 1 |. are presented these organs in situ, as
they appear viewed from the right side, both the valves having been removed.

Papilio Lycidas, Cram. (Plate XXXII. figs. 15, 16.)

The valve is trigonal, short and wide, the margins fringed with black shining hairs
mostly springing from within ; the interior surface dark brown, velvety, nearly flat to a
considerable distance within the margins, then suddenly sinking to a deep semicircular
cavity. Within this hollow is seated a short harpe, having a widely expanded base,
sloping away to the ventral side, and soon rising to a free flat column of dark-red glit-
tering chitine, which is produced at the two corners into long sharp spines. Besides
these, there are two minor spines from the slanting dorsal edge, and one near the base in
the ventral. I discern no trace of serration in any part.

The abdominal apparatus seems peculiar. The beautiful clothing-scales, velvet-black
alternating with pale-yellow, project densely from the segment around the cavity, when
a valve is removed, so as greatly to conceal the interior. My cabinet possesses but a
single example of the species, which is both beautiful and rare ; and this I was reluctant
to destroy, contenting myself with close and long-continued study of such organs
situ as a powerful hand-lens could resolve. Conspicuous is a slender well-formed
scaphium, projecting horizontally, of a shining brown hue, with the " double teeth " small
and low, but duly developed. At first there seemed no trace of an uncus ; but at length
I found it very far back, under the thatching roof-scales, small and short, sloping
downward from its oiigin, and presently bent with an abrupt angle to a quite vertical
line, or even a trifle within the vertical. Its form and direction, as well as I could
see it, arc much as in P. Mector. The uncus, then, it is evident, can here present no

IN CEKTALN LEPIDOPTEEA. 323

antagonism to the projecting spines of the harpe. These points, in the closed condition of
the valves, meet at the very tip of the scaphium ; yet the scaphial teeth, which in some
species are long and formidable spines, are here minute and apparently useless.

To complete the series of anomalies, I could find no trace of a penis. By so holding
the abdomen that the nearly level sun's rays shone into the cavity, I once or twice
detected a glittering point very far within ; whether this was a penis, or whether the organ
had been accidentally lost in my preparatory manipulation of the specimen, I dare not say.

Papilio Doubledayi, Wall. (Plate XXXII. figs. 17, IS.)

Va/re having the outline of a tall cone, edged by a very broad fringe of close-set hair-
scales — yellow at then' bases, and becoming pale scarlet at their tips — which are con-
tinued from the exterior surface.

The harpe is commensurate with the valve itself, from which I am not able to
separate it. The extremity is double, the dorsal point continuing the general plane ; the
ventral really much longer, though not apparently, because it bends upward in the line
of vision ; it is fringed with fine short diverging hairs, proper to its point (i. e. not
continuous with the exterior clothing). The dorsal edge also bends up, and becomes
semi-erect ; its outline is so cut that, in the middle, it forms a sharp tooth, pointing
backward. Except this, I find no serrations in any part of the structure. The pre-
hensile function seems to be limited to the upcurved dorsal point, and to be feeble.

The uncus and its accompaniments are exactly as in P. Rhodifer (infra), even to the
protrusion of the -wiry penis, with the single exception that the scaphium is here deeply
keeled, whereas the keel seems totally absorbed in P. Rhodifer.

I subsequently examined another example, in which the penis was exactly as in
P. Rhodifer, except that the finger-point was not so produced, and the lower lip-like lobe
was rounded and of a deep red hue ; but the extruded bladder described there was here
a perfect little globe of translucent yellow substance, projecting like a bubble from the
expanding lip.

A third example had the tip of the very same form, but without a trace of the vesicle,
the hollows of the bifid extremity white and shining.

Papilio Phodifek, Moore. (A. D.) (Plate XXXII. figs. 19-21.)

In the valve of this Andaman representative of the Coon group there is a marked
generic resemblance to that of P. Doubledayi, together with ample specific difference.
The description of the valve of Doubledayi will well suit this of Rhodifer, save that
the fringe, which is even broader, is of a purer scarlet hue, and the outline is still more
disguised by the dorsal projections.

The harpe is again commensurate with the valve, and inseparable from it by mecha-
nical lifting ; though, as in Doubledayi, the brilliant glittering polish of the entire surface,
which distinguishes the harpe proper from the dull lining-membrane of the cavity, shows
in each case that it is organically distinct from the valve. Here the extremity is treble ;
the dorsal point, much larger, and directed much more laterally, than in Doubledayi,
stands erect; while the ventral division sends off from its surface, a little below its tip,

•V'

324 ME. P. II. GOSSE ON THE CLASPING-OEGANS

an adventitious lamina, of conical outline, which also stands up in the line of vision, and
doubtless constitutes an auxiliary prehensor. All the three points are thickly beset with
fine hairs, which make the resolution of their form difficult, and, one would think, must
interfere with their prehensile effectiveness. There is no secondary tooth on the dorsal
edge, which curves inward in a deep semicircular sweep to the broad expanded base.

The uncus is very long, and of excessive slenderness, almost quite straight, its ex-
tremity accurately reaching the point where the dorsal elevations of the two harpes meet
and touch each other. The scaphiurn is very small and shallow; its " double teeth"
conspicuous as a great spine on each side, having a downward curve. The most remark-
able feature is the penis, which, like the uncus, is of extraordinary length and slender-
ness, resembling an attenuated, highly polished, black wire, nearly straight, proceeding
from its conical investment almost directly downwards, and actually protruding between
the valves below to the distance of an eighth of an inch; so that the insect could be
easily handled and turned about by it, as if it were an inserted pin. The extremity of
this organ is produced above (structurally above, though actually beneath, because its
position is reversed) into a finger-point, which is abnormally long and acute ; while
below it forms a wide, somewhat tumid, trumpet-lip, of exceedingly attenuate trans-
parent yellow chitine (?). So far the appearance has been paralleled in many other ex-
amples of the organ ; but what follows is more remarkable. Erom out of the trumpet-like
orifice proceeded what seemed a blown bladder, wider than the orifice itself, as if it had
expanded as it issued, not quite globular, of a substance resembling that of the lip, but
still more delicately thin, which, however, had dried in form and in place. It had all
the appearance of an integral organic tissue — not of an excretion.

Papilio Akchesilaus, Feld. (Plate XXXII. figs. 22-24.)

J'alce rotundo-triangular ; the extremity very obtuse, but fringed with graduated hair-
scales projecting into a long and sharp point ; the other margins only minutely fringed.
The middle of the cavity is occupied by a large harpe, crescentic when viewed vertically
to the plane of the valve ; the cusps, which point ventrally, are raised considerably from
the floor ; the one next the base pointed, the other obliquely truncate. Both the points,
and the interior curvature, as well as a ridge that takes a parallel curve, are serrated
with sharp teeth.

The uncus is horizontal in its total direction, but has a double sigmoid curve; and its
extremity, which is suddenly bent doAvnward, is trifid, with blunt points. The scaphium
is small, and so closely adherent to the uncus that the boundaries can scarcely be dis-
cerned. I detect no trace of the " double teeth." The penis presents nothing specially
notable : it is small, fingered, and retreating.

Papilio Zalmoxis, Hewits. (Plate XXXII. figs. 25-28.)

Valve large, trigonal, with rounded angles, nearly equal-sided. Exterior densely
clothed with shaggy fur of hair-scales, which project from all the free margins, forming
an even, wide, buff fringe to the interior side, which is deeply concave, almost hemisphe-

IN CERTAIN LEPIDOPTEEA. ,325

rical. From the dorsal margin rises a wall-like ridge of shaggy surface and summit,
which inclines towards the concavity, curves around the point, and then ceases. From a
broad conical eminence at the base (the attachment to the basal knob) rises the harpe,
a plate of chitine nearly filling the concave floor, its ventral edge elevated into a ridge,
which curves in a semicircle to the dorsal wall ; the middle of this ridge is lengthened
into a long polished black curved spine, broad at its foot, becoming slender towards its
tip, which projects over the fringe of scales, just reaching its extremity. At first sight
the point of this spine appears simple ; but higher magnifying power shows that the
extreme point is flattened, and cut into four very minute transverse teeth, of which one
is longer and larger than the rest.

The uncus is reduced to a small horizontal projection of the tegumen, bluntly pointed,
of polished brown chitine. the lateral rami rising considerably above the level of its
median ridge, the midst of which is studded with hair arching outwards. Depending
from it is a sufficiently ample scaphium, of which the two sides slope downward and out-
ward, and the keel descends from the angle. Xo distinct " double teeth " are developed.
The colour of this organ, usually white, was, in two examples examined, of a dull blackish
brown ; and the surface was shrivelled and corrugated, not symmetrically, and not alike in
the two examples. The penis is long and slender, slightly enlarged at the tip, black,
polished ; in one example it reached almost the edge of the valve-fringes when closed.

Papilio Policknks, Cram. (Plate XXXIII. tig. 1.)

The valce is one of the most aberrant that I have seen. It is parallel-sided, the end a
segment of a circle with a projecting point ; exteriorly it is very densely clad with long
scales and hairs, the dorsal half black, the ventral white, the colours abruptly divided.
The interior aspect is affected by this clothing; all up the ventral side the white furry
coat forms a projecting back-ground to the valve, increasing around the extremity, till it
becomes nearly half as long as the valve itself. But up the dorsal side the outline is
quite lost in close black hair, which, at this quarter of the end, becomes suddenly a thick
tuft longer than the valve.

The harpe is very complex, aud hard to define, harder still to describe in words. Near
the tip there are three transverse rows of strong teeth : — first, one tooth at the very
edge, making the point, and another at its dorsal side; below a ridge runs across,
bearing six stout and prominent teeth ; and below this, from the dorsal margin, arches
out a thick arm bearing a sort of open hand or curved claw, of four fingers. All of these
stand up from the cavity, and project inward. Each ridge, moreover, sends down one or
more longitudinal ridges, more or less jutting into points. The whole constitutes a very

I elaborate prehensile instrument, quite unique, so far as I have seen.
The uncus is also abnormal. It is very short, broad, rounded at the tip, with slight
curvature, of pale yellow chitine. I could not detect any sign of the scaphium ; but my
observation was unsatisfactory.

Papllio Ucalegon, Hew. (A. D.) (Plate XXXIII. figs. 2, 3.)

Valve very small, parallel-sided ; the extremity round, with a regular wart-like pro-

326

ME. P. H. GOSSE ON THE CLA SPING-ORGANS

iection at the tip, which is darker and seems firmer and denser than the substance of the
valve as if it partook of the nature of a prehensor. The dorsal margin is fringed by
very Ion- projecting hairs; and the interior cavity is thickly beset with fine sinning
pale hairs, in irregular groups, which spring from various points, and make it difficult
to see the'outhne of the parts. In the delineation, I have omitted most of these tufts of
hair as otherwise nothing could be represented. The harpe is complex. Pirst, from the
dorsal margin, a little below the projecting terminal wart, rises, almost erect, a stout
black polisned spine; then, further down the same margin, but a little within, is a
similar smaller spine of clear yellow chitine. From a base occupying the whole width
of the valve, proceeds along the middle of the cavity an irregularly curving, stout, thick
rod whence, beyond the middle, springs an erect, black, polished spine; the rod then
expands, and terminates in a free claw of two pincers, notched along their opposing

g(J(TgS.

The whole anal region is protected and concealed by a wide canopy of long, pro-
jecting, parallel scale-hairs, over all, which reaches far beyond the fringes of the valves.
These proceed from the dorsal arch of the eighth segment ; and other long scales proceed
from the sides of the same. When these are all removed, we see a moderately long
uncus, very slender, acute, nearly straight, and horizontal. Below it is the scaphium,
dark but translucent, shining yellow, shallow, and narrow but long, the " double
teeth" discernible only as a slight thickening.

The penis is very long, narrow, pointed, in contact with the lower surface of the
ecaphium, beyond whose extremity it protrudes.

This and the preceding, both African species, have, in their irregular and complex
armature, much in common, with ample diversity.

Papilio Agesilatts, Boisd. (A. B.) (Plate XXXIII. figs. 4-6.)

Valve externally thickly clothed with scales, half white, half black, the colours abruptly

divided Along the ventral margin, around the extremity, and nearly halfway down,^

the dorsal side, a close fringe of long black hairs extends, springing from the very edge.

This considerably augments the apparent area. The cavity is almost wholly occupied

by a large and very elaborate harpe, whose structure I do not quite understand; for

thou-h ft can be separated for examination, the process inevitably distorts the parts,

and alters their relation to each other. There seems to be a great arching ridge of dark

chitine right across the middle, ending in a great erect double claw, near the ventral

margin'- more or less structurally connected with this, by the chitinous base, rise two

other erect pieces, the three standing triangularly, and bending toward each other.

Prom the arching ridge extends a (comparatively) great spoon-like plate, nearly filling

the terminal moiety of the valve, its ventral edge, running about parallel with the

ventral edge of the valve, of denser and darker chitine than the rest, cut into close

minute saw-teeth. This saw-like edge stands, throughout, free above the level of thd

""In examination of the left valve in some measure cleared the difficulty— at least so

IN CERTAIN EEPIDOPTERA. 327

far as to show that what I have called the spoon-like plate is not, as looked at, concave,
but convex, the outside of a hollow cone or helmet.

The tegurnen, in situ, is concealed under the seventh segment, even when this has
been considerably denuded ; the uncus, however, protrudes, but much disguised. For
it is almost straight, projecting horizontally, with a slight sigmoid curve, and having
for its termination, instead of the usual spathulate point, three short, slightly divergent
fingers. Beneath this representative of the uncus there is the long narrow scapkium,
with shallow keel ; but, from the close contiguity of the parts, and their minuteness, I
could not be quite certain of the presence of the " double teeth," though once or twice
I thought I discerned them.

If we compare this fine American butterfly with the still finer, closely related P. Arclie-
silaus of Felder (p. 324 supra), we find almost identity in the form of the uncus, but
great divergence in the forms of the valve and the harpe.

Papilio Papmatus, Gray. (A. D.) (Plate XXXIII. fig. 7.)

This Australian cousin of the preceding species has a valve and a harpe equally com-
plex, and equally hard to demonstrate ; while the details of the armature seem altogether
its own.

The valve is unusually small; and the presence of many irregular tufts of close long
hairs, in unwonted places, greatly conceals and disguises the structure.

The harpe is, as I have said, very complicated, composed of many pieces, whose forms
and positions and directions seem to have no intelligible relation to each other, or to
the common object. There appears to run around the terminal portion of the valve a
slender chitinous framework, which has two, if not three, projections parallel to each
other, one much slenderer and longer than the other, forming a free arch. Down the
dorsal side is a thin Avail, rising in the middle into a tall, thin, erect, curved tooth,
directed backwards. On the ventral side a ridge runs, within which, near the middle,
is a semiglobular plate of deep-brown chitine, having two tall erect incurved faces,
the inner forming a simple broad tooth, the outer and opposite cut into many acute
serratures.

The abdominal apparatus, so far as I have been able to resolve it, seems to consist
of a very small and compact scapkium, with the uncus (ill-defined) lying close upon it,
and a penis moderately produced, reversed, terminating in a swollen extremity and a
curved fin ger-p oint .

These very elegant butterflies appear emphatically and characteristically PapiMones ;
yet, in their prehensile armature, they show a manifest approach to those extra-Papi-
lionidan families in which the valve and harpe are united to form a single complex
prehensor.

Papilio Codrls, Cram. (Plate XXXIII. figs. 8-12.)

Valve parallel-sided, circular-ended, fringed with very long, badger-grey, shining
hairs. A broad shelf runs obliquely across the end, and narrows down the sides. Near

328 ME, P. H. GOSSE ON THE OLASPING-OEGANS

the median lino of the cavity, from a base that spreads like the base of a tree-
trunk, runs up a narrow harpe, nearly to the arch, and then bends over towards the
dorsal margin. It consists, on the ventral side, of three parallel lamina?, divided by
deep sulci, but united beneath, where the whole is in union with the lining-mem-
brane. Of these lamina?, the one next the ventral margin rises, abruptly, near its
middle, to a higher level than the others, which it maintains. All three have the free
edges cut into saw-teeth, on the ventral coarsely, on the others finely.

The abdominal apparatus is very abnormal. I cannot find any trace of an uncus, nor
of a scaphium — unless, indeed, a minute rounded projection clothed with white hairs,
in the usual place of these organs, represents either or both of them. Then, the penis
appears to be double or even treble ; for after I had removed, for examination, one
organ, which corresponded, in position and form, to the penis*, there remained another
which had been beside it ; and this consisted of two parts — a hard chitinous thin blade
which terminated in a slender spine, and a membranous investiture which ran up its side
half embracing it, separating near the tip, and forming a thin yellow finger-like lobe.

All these organs, however, need fuller examination. The termination of the abdomen
is profusely studded with long crowded grey hairs, under which they are almost hopelessly
concealed ; they are, moreover, very minute. The specimen examined was the only
example of the species in my cabinet. I have represented what I saw ; but I am not satis-
fied. The species is manifestly very anomalous.

Papilio Nireus, Linn. (A. D.) (Plate XXXIII. figs. 13-10.)

Valve approaching a trigonal form; scales of exterior slightly fringing the edge.
Cavity moderately deep ; a fringe of long hairs springs from the whole length of the
dorsal margin, and arches over the interior.

Harpe a flattened rod or narrow plate, perhaps hollow, springing from an expanding
base, and following the middle of the floor to the extremity, where it becomes spathulate
(PL XXXIII. fig. 13). It is with case removed entirely from the valve, when we
see that the spoon-like tip, which is sometimes flat, sometimes hollowed, is studded with
short stout spines, closely set over its surface (fig. 15), and that the dilated base rises
into two or three conical eminences, one of which is cut into saw-teeth along its dorsal
edge (fig. 1-1).

The uncus is an exceedingly slender spine with a fine-drawn point, projecting nearly
horizontally, with the slightest possible curve downward. It is keeled to about mid-
length. The scaphium is much more conspicuous, bent obliquely downward, with two
stout, blunt, erect, chitinous processes, which appear to represent the normal "double
teeth." The penis has a long finger at its upper extremity. In situ, all these parts are
concealed by the coarse shaggy coat of scales with which the abdominal segments arc
clad f.

■■' I regret that I can say no more than this; but, when I had laid this organ on a glass slide for examination,
a draught of air carried it away, and I could not find it again.
t Sec remarks on P. Bromhis, at page 320, supra.

IX CEETAIN LEPIDOPTEEA. 329

Papilio Diphilus, Esp. (Plate XXXIII. figs. 17-20.)

(1. The larger race, figured by Esper, Aus. Sell. tab. xl.)

The genitalia are developed under peculiar conditions here. Looking at the d , we see
at first no appearance of the usual valves, and the whole aspect is abnormal. The eighth
segment does not protrude from under the seventh, but sends out, just at the edge of the
latter, a dense arching thatch of long, horizontal, scarlet hair-scales (see PL XXXIII.
fig. 20), under which the genital organs are widely exposed, much as in JEurycus and
Euryades.

The valves must be searched for, and then will not be easily recognized. Each is very
small, not more than -05 inch in total length, not clad with scales externally ; indeed, I
think it would be correct to say, there is no valve at all, but only a naked harpe of valve-
shape. Its outline is that of a hand-bell; the margins are thickened, and united at the
summit into a rounded knob (=the handle of the bell), which is beset with minute curved
sjDines, and bends-over towards the dorsal side *.

The uncus is a nearly straight wide process, semitubiform, the extremity truncate, but
cut into seven obtuse teeth. The scaphium is seen beneath, narrow, short, with only a
suggestion of the usual " double teeth." The penis is more normal, wiry, polished, acute,
projecting beyond the surrounding organs, slightly above the horizontal.

(2. The small race.)

I received a variety from Malacca, much smaller and more elegant than the Java type
as figured by Esper, pi. xl., and by Cramer, pi. csxxviii. ; the wings longer and narrower
in proportion, the colours much brighter, and the white spot of the macular band in the
hind wings longer and narrower, with other slight differences. In this, the bell-shaped
valveless harpe is exactly as in the larger form, and is of the very same dimensions.

The uncus appears in this variety somewhat more truncate, shorter and less cut into
points, so far as I have been able to examine it ; the scaphium more keeled, and the
other accessories the same. There is a pair of organs forming the floor of the anal cavity,
with prominent points, the edges beset with long hairs, which I at first thought might
possibly represent the ordinary valves ; but so shallow that the harpes are separate from,
and project above, them. But these are possibly the pair of strong angular projections to
which the expanded bases of the harpes are usually attached. Their high position here
agrees with the position of the harpes, and confirms my conclusion of their nature.

Papilio Polydokus, Linn. (Plate XXXIII. figs. 21-23.)

There is much resemblance between the terminal region of the body in this species
and that in P. Diphilus, as might be anticipated ; but the structure is here even more
aberrant. The horizontally projecting thatch of scarlet hairs is still more copious,
descending much further on each side, and even continuing, in coarser form and more
irregular arrangement to the bottom. Thus, though the shining black genitalia can be
seen, nothing can be defined till this shaggy thatch is removed ; which is accordingly
done in my sketch of the parts in situ (PI. XXXIII. fig. 21.) Probably the protecting

* Compare it with fig. 15 of Nircus.
SECOND SERIES. — ZOOLOGY, VOL. II. 46

330 ME. P. H. GOSSE OX THE CLASPING-OEGANS

office of the ordinary valves is to some extent supplied by this thatching, the valves
being almost aborted.

The uncus is here a stout, short, bifurcate process, the two points not quite equal
inter sp, acute, notched below. They are followed by a couple of pear-shaped organs,
of highly polished, glittering, brown chitine, each terminating in an oblong knob, beset
with minute spinous teeth. In fig. 21, they are seen in .s'dit ; in fig. 22 the right is seen
interiorly, showing that it consists of a hollow shelly ease, with the edges curved-over.
The knob bends upward, toward the bifurcate uncus, but comes considerably short of it (as
shown in fig. 21, where both the right and left knobs are seen), whereas in P. Diplal/ts
the knobs close on the truncate tip of the uncus.

There can hardly be a doubt that these organs are homologous with what I have called
"valveless harpes," in Diphilus. But, tracing them to their origin, we find that they
spring from a point on each side, within the descending rami of the uncus, i. e. within
the walls of the ninth segment; whereas the valves are appendages of the eighth, and
are always outside the walls of the uncus *. Thus it is perhaps possible to consider these
spinous knobbed organs the homologues of the teeth of the scaphium, which (in P.
Merope) we have seen to be serrated. The scaphium itself I do not find in P. Polydorus,
unless it is a shining white mass (in texture like the white of an egg boiled hard), into
which the pear-shaped organs merge at their bases.

What I suppose are the col res proper (or their representatives), are a pair of organs
(seen /'// situ, in fig. 21 below the pear-shaped bodies, with the penis protruding from
between them) which certainly spring from the bottom of the eighth segment. Fig. 23
represents one of them, the left, viewed internally ; it bears a semitubular ridge of chitine
running through its midst, terminating in a minute point. This (supposed) valve is
hollow, and is beset, on both margins, with long straight bristles.

If this analogy is true, it must be true also in P. Diphilus ; but the application seems
more difficult there.

It may be interesting to compare De Haan's description with mine above, every
word of which, I may say, was written before I saw his. He says, "P. Liris and P.
Polydorus have the hinder part of the body naked; the valves [kleppen] are very
short, armed from within to the top by a hook [?, P. II. G.], and the lateral appendages
[zijdelingsche aanhangsels] of Amphrisius t are placed under it : the spine on the back
is blunt [uncus ? P. II. G.] ; besides this, there are also two spines present on the
upper edge, and these are curved inwards ; so that, altogether, the outer edge is provided
with seven appendages J, which all rise free from one another. Between these lies the
penis [roede], which, as if pressed together, has a sharp edge above and below "§.

* The harpe however, has a double basis normally first to the valve, secondly to the knob which projects from the
floor of the genital cavity. This latter is, I think, situate within the ninth segment.

t These are, I presume, what I have called the harpes in Ibis Memoir.

J If I understand him, I can find but six : viz. the two points of the bifid uncus, the valveless harpes (or sca-
phial teeth) ; and the two bristled valves? Hut I am not sure that I follow his identifications. He seems to imply a
terminal point to the uncus, besides and between the two points which alone I find in Polydorus ; and in his figures
of Liris he represents such a point. Polydorus he has not figured. Possibly he did not examine both minutely.

§ Nat. Gesch. Ned. 0. B., Tapilionidea, p. 17.

IN CEETAIN LEP1D0PTEEA. 331

Papilio Antenor, Dru. (Plate XXXIII. figs. 21-26.)

Valve of the rotundo-triangular outline common to tlie genus Oru'dhoptera, and nearly
as large as in some species of that genus. Exteriorly of an uniform drah-white ; a fringe,
of dense hair-scales, projects beyond the dorsal margin, the continuation of the common
clothing ; the ventral margin is unfringed. Interiorly, the lining-membrane is of a
rich umber-brown hue, dully shining ; the contour forming a wide flat edge or shelf all
round, whence it falls abruptly in a deep cavity. This is occupied by a wide bell-shaped
harpe of glossy chitine, hollow, like the cavity which it occupies, and terminating in a
rounded knob, slightly bent towards the dorsal side. This is almost exactly the figure
and appearance of the (presumed) harpe in Diphilus, only that I cannot detect the slightest
trace of serration on the knob or on any other part. The knob is spoon-shaped, hollow,
with an elevated, thickened, overturned edge.

The uncus is nearly horizontal, straight, short, bluntly pointed, being thickened at
the edges and tip. The surface, though shining, is uneven, and lacks that glittering
polish which this organ usually displays. A shallow sulcus runs along the summit
to the point. Below, attached to it by short white fibres, is a small scaphium, in which
I detect no teeth. The penis is very minute, and scarcely visible, far within the abdo-
minal recess.

Papilio Hector, Linn. (Plate XXXIII. figs. 27-31.)

We have here the same type of structure as in Polydorus, but with remarkable modi-
fications of the organs in detail. There are, again, no proper valves ; the posterior
extremity appears as an oval opening ; the area occupied by a confused set of shining
black knobs, surrounded by a dense horizontal thatch of scarlet hair, which, however, is
not so long as in Polydorus.

The uncus is a short spathulate point, abruptly bent perpendicularly from the middle
of the transverse edge of the downsloped roof of the ninth segment. Its shape, as seen
directly from behind, is shown at fig. 28.

At its point just meet the summits of a pair of organs, answering to the (suggested)
scaphial teeth (see Polydorus, p. 330), but of very bizarre form. Side wise they an1
not unlike the corresponding organs in Polydorus, very highly polished, like them ;
but when looked at directly from below, each is seen to send off a broad process of sub-
trigonal outline, horizontally, toward the middle of the cavity. The process is a sort of
flat cushion, apparently of short close pile, throughout which are set minute slender
needle-like spines vertically. The end is a thickened knob, whence descends obliquely
a group of very long straight bristles (see fig. 29, the right "scaphial tooth" seen
from below).

What I suppose to answer to the valves, are still further disguised here, each being
a slender straight rod (fig. 30), with the point bent dorsally, nearly at a right angle, and
bearing a short spine ; while the dorsal margin is so beset with short straight bristles, as
to give to the whole organ the appearance of a comb. The base is broadly dilated ; and
the pair incline towards each other, so that the summits are almost in contact.

Between them projects the very much developed penis, the tip forming a curved acute

46*

332 ME. P. H. GOSSE ON THE CLA SPING-OEGANS

spine, reaching beyond the level of all the other organs, and the preputial (?) appanage
forming two ample parallel leaf-like expansions, greatly lengthened, situate on the dorsal
side (fig. 31).

Professor TVcstwood, who figures this showy species in 'Arcana Entom.' i. pi. 3, mentions
the affinity shown by its larva and pupa with those of P. Polydorus, on the authority of
Horsfield and De Haan. And, what is much more curious and inexplicable, there is a
very close resemblance between all * these organs in P. Hector (including, notably, the
bent bristled expansions) and the corresponding organs in a great Asiatic Silk-moth
(Anthea Roylei).

Claspixg-Orgaxs ix Butterflies or other Families.

I subjoin a few notes made on species of other Families, chiefly the Pieridce, because
it is interesting to trace the gradual disappearance of a vanishing organ or set of
organs.

Pieris. The valve of P. Autodice, of South America, a very charming microscopic
object, densely clad in its snow-white hair-scales, is essentially that of a Papilio. The
harpe is a thin sharp knife-edge, colourless, ulassy, running transversely across, not far
from the base, and rising into a broad tooth in the middle. The tegumen has the bird's
beak-form; but I see no development of a scaphium. The penis is large and prominent.

The valve of P. Automate is much the same. A harpe runs obliquely from the base
to the upper part of the dorsal edge — a colourless glassy ridge, with a free point. The
valve, in both of these two species, is armed with a projecting semicircular hook on its
margin, at, or close to, the tip.

In P. Papce the uncus is well developed ; but I find no trace of a harpe, nor any structure
corresponding to it or displacing it.

P. Brass'icce possesses a well-formed valve, whose outline may be described as a circle
within the angle of two sides of a square, of which one is the base. The other side is the
ventral, which runs up with a stilf straight edge, and terminates in a hooked point. As
harpe, an acute spine of transparent chitine runs from the base nearly to the middle,
and then bends up toward the dorsal side. Herold takes no notice of this little glassy
rape.

Callidryas. C. Eubule has a very curious valve, armed as elaborately, and as singularly,
as that of many a Papilio. Its outline is somewhat pear-shaped, having a large hook at
the extremity, and a broad shoulder on the dorsal side, each of which rises into an
upturned black chitinous tooth. Through the middle of the valve runs a wide and deep
depression, in which the floor is so thin as to be translucent. In the very centre there
is a curious, oblong, free body, of orange hue, studded with a score of erect, blunt, black
spines, equally thick throughout, looking like needles carelessly stuck in a pincushion, or,
like the shell of a sea-urchin ; it is connected, by a projecting arm, with the dorsal margin.
There is a beak-like uncus, little hooked, beneath which projects a conical scaphium,

* Except the uncus. The form of the tegurncn in A. Roylei maybe imagined by supposing a normal uncus, only
with the terminal half divided into two divergent spines, and the interval between these tilled up, to a line stretched
from point to point, with a downy chitinous membrane, and then bent down.

In Polydorus, these points remain bifurcate, as they are in the American Moon-moth (Aetias Luna).

IN CERTAIN LEPIDOPTERA. 333

transparent, and of extreme tenuity. The penis is of unusual length and slenderness,
resembling a very fine copper wire.

In 0. Statira, the valve is still more remarkable. Here the pear-shape is lost ; for the
extremity runs into a long, sharp, curved spine, and the dorsal shoulder has two wide
semicircular excavations, on the eminence between which stands, attached by one angle,
a triangular plate, the outer edge of which is notched into a dozen saw-teetb. This, I
presume, is a harpe, and replaces the pincushion of G.Eubule. The tegumen projects
in along sballow channel, which can hardly be called an uncus, though it replaces one ;
beneath this is seen a minute scaphium, which simulates the form of an uncus. Tbc
penis is long, arching, and wiry, much as in Eitb

Goneptenjx. Tbe valve of G. Ehamni is small, but well made, rotundo-triaugular,
oblique, terminating in three horny points, of which the dorsal one is a rather long,
slender, acute spine, which leans across the valve, so as to project beyond the ventral
edge. This represents, but is not, a proper harpe. There is a small well-formed uncus,
which, strangely, is translucent white, and a large penis, but no trace of a scaphium.

In the fine G. Leachktna of Brazil the valve is much like this, with very wide fringes
of yellow hair-scales. The terminal point is long, slender, upcurved ; and the two dorsal
points both belong to that which supplants a harpe. There is a proper uncus, small,
hooked, and a scaphium, very minute but apparently normal, beneath it. The penis, in
this species, is long and prominent ; its terminal expansion takes the singular form
of two triangular lobes, looking exactly like the anal and dorsal fins of a mullet.

G. Clorinde, of Paraguay, has a similar valve. The uncus is a large blunt hook,
polished white, like ivory, which bears beneath it a perfect little scaphium, in which,
I believe, I can trace the " double teeth," though excessively minute.

Hebomoia. H. Glaucippe is well armed. The valve is an oblique trapezoid, ending
in two stiff slender spines ; the ventral margin semioval in outline, furnished with a
' broad white fringe. Its long very slender harpe runs transversely across the base,
! which throws up an elevated tooth at the dorsal end, and terminates ventrally in a more
or less produced hook. It is quite papilionine in aspect, though no part, I think, is
serrated. There is a long black uncus nearly horizontal, but bent down at the point.
This is furnished with an ample scaphium of a polished brown hue, quite dark in parts .
it has a long keel, projecting beyond the hook of the uncus, with long and deep rami,
and a stout, horizontal, decurved black tooth on each side.

Tei'ias. The valve of T. Ilecabe is semicircular, running off to a tooth-like point at
the vertex. The concavity is deep, almost hemispheric ; and the margins are very broadly
turned over, their inward edges elaborately furnished with spines, long, slender, sharp,
arching over the cavity. On the dorsal side of the terminal tooth rise two of these spines,
and a third on the ventral side, while the ventral point of the basal side runs off into a

I stout and sharp hooked point. There seems to be no harpe.
The uncus may easily escape recognition. Not only is it exceedingly minute, but it
seems to be replaced by a projection with two points, the upper curved downward, the
lower, longer, curved upward; but the former is the uucus, the latter its scaphium.

334 ME. P. H. GOSSE ON THE CLASPJNG-OEGANS

Colias. Here the genital armature is become very aberrant. C. Edusa bas a short,
tbick, obtuse uncus, and no valves nor barpcs, but fixed side pieces in place of valves, in
wbicb I can detect little analogy with Papilio.

Mbrpho. In M. Menelaus I find valves mucb as in Papilio but witbout barpcs, an
uncus broadly triangular but witb a good booked beak-like point, very deeply keeled, and
a projecting penis. Tbe tootb on eacb ramus of tbe uncus is more tban usually separate
and strong. I see no trace of a scapbium.

Nor in 31. Epistrophis, in wbicb for a while I fancied I did see a rudiment of tbe organ
far up in tbe cavity ; but very precise and repeated observation under a bigb power, with
ligbt thrown well into the spot, convinced me that I was looking into tbe anal orifice, and
that there was not tbe least trace of a scapbium.

In tbis species there are well-developed valves, but small, and of delicate texture, of
which the tip, preceded by a pointed fringe of very long silver hairs, is of hard, dark
ebitine, turned up at an angle, with a serrated edge, of which the teeth are large, and cut
into minute toothlets along their sides. Here, then, we have one of the most peculiar
characteristics of a harpe transferred to the valve.

Dynastor. But in the fine I). Barms the transition is complete. Here is a narrow,
parallel-sided but very deep uncus, and a pair of horny stout organs in the place of
valves, and having mucb of their outline, but bearing along the thickened dorsal margin
a double row of strong spinous teeth, black and glittering, most formidable to behold.
The valve with its harpe has become what Dr. White would call a true harpago, distantly
like his beautiful figure of the organ in Argyrmis Laodice (op. cit. pi. lv. fig. 20).

I now close this Memoir ; but the work is very far from complete. It is but an
instalment of tbe subject that I here present to the Linnean Society ; a great majority
of tbe recognized species of Papilionida? being still untouched for this purpose. Yet I
have not ceased to work at them ; and if life and health be still vouchsafed to me,
I hope I may, before long, contribute to science more details of these remarkable instru-
ments, as yet unsuspected*.

Conscious I am, too, that a good deal of imperfection attaches to tbe observations
recorded. Of not a few of the species examined and figured herein, I should be very glad
to repeat tbe examination, if I bad more specimens ; and Entomologists could scarcely
render me a kinder or more grateful service than by sending me examples of male
Papiliones, however torn in the wings, or even the separated abdomens, if duly authen-
ticated, of which they may possess worthless duplicates.

* At the time of this sheet going to press (March 6, 1883), I have accumulated, in MS., descriptions, with drawings, ■
of the genitalia of fifty-six speeies, in addition to those herein contained, belonging to the genera Orniihoptera, Papilio,
J\ inopalpws, Sericinus, and Leptocircus.

IN" CEETAIX LEPIDOPTEEA. 335

APPENDIX.

Supplementary Note to Ornithoptera Remus.

Since this Memoir was completed, and presented, many examples of this magnificent
butterfly have been sent to me from Celebes, and also of O. Haliphron ; and I have
examined more than a dozen males of each. The following notes embody my latest
researches on the abdominal organs. They are illustrated by figs. 15-19 of Plate
XXVII. [N.B. The small letters indicate the same organ or part in each figure.]

Uncus (a). When the beautiful patch of velvet-black hair-scales that adorns the
dorsum of the eighth segment of O. Remus is carefully removed by abrasion, so as to expose
the chitinous skin, we see that this is dully shining up to the accurately defined trans-
verse line where the velvet ended. Thence it becomes brilliantly polished, and begins
to project into a triangular area, of which the lateral edges are thickened, while the
median portion, also thickened, forms the out- and downcurved uncus. Its curvature
is not quite uniform : sometimes it is the arc of a circle, sometimes the arc of an ovoid ;
sometimes it is bent abruptly and perpendicularly downward from its middle.

Scaphium (b-i). From a level slightly below the expanding rami of the uncus, appa-
rently in fixed immovable connection with it, spring a pair of slender rods, of similar
shining chitine, each of which (b) is dilated near its base into a wide lamina exteriorly,
which lamina? become the firm fleshy tissue of the cheeks (c). These together assume
the form of one fourth of a globe, cleft by a deep vertical sulcus, which has a continuous
floor (d). Each cheek has its outer ventricose side obliquely cut away behind : its
surface is distinctly granulate, the granules elevated, more and more towards the point
and back, into the characteristic stiff, glittering bristles (e).

A simple device may assist the comprehension of this form. If the thick rind of half
an orange be removed bodily and set on its edge, if then about half of this be cut away
slantingly behind, and then what remains lie excised into a deep narrow groove mesi-
ally, we shall get the scaphium-cheeks of O. Remus.

Erom just below the groove descends vertically a long and narrow pentagonal piece

J (/) of what simulates polished ivory. It is not continuous with the part from which it

i seems to descend ; for, by peering with a lens under I lie edges of the cheeks, we see the

summit of the pentagon to lie abruptly bent backwards, so as to reach connection with

the under surface of the sulcus-floor, far back in the mid-roof of the arch.

Behind this narrow shining pentagon is the keel {g) ; a large plate of whitish tissue,
I very thin in the middle, which seems attached to the concealed side of the pentagon, but
on the abdominal side thickening into broad irregular lobes (/i), apparently of a different
t character from the thin central parts. The depth of this organ, and even its shape,
differs considerably in different individuals. Possibly it is capable of alteration in form,
and may be composed of something analogous to erectile tissue. Indeed, I have fancied
that the varying direction of this curious member — sometimes more, sometimes less
pendent — may itself indicate a sort of erection. May it be that it is a titillant, excitant,

336 ME. P. H. GOSSE ON THE CLASPING-ORGANS

provocant organ, to be applied to the ? vagina, before intromission ? The extreme
tip seems to be of peculiar structure : it may be glandular, and bo the seat of sensation,
the reW-c epvriKv, of wbich the hard chitine of the penis seems incapable.

In one example the whole keel was movable, as on a hinge in the interior, whose
place is indicated by the dotted curve-line in fig. 15.

The thickened edges unite behind into a very thin median ridge, which, like a groin
from a roof, descends vertically, and is lost in the curtain to be presently described. A
transverse horizontal section would be somewhat like fig. 19.

In O Remus, and, I believe, in Ornithoptera generally, the uncus is soldered by its
under surface to the upper surface of the scaphium-cheeks behind, the suture showing a
thickened bead. So it is with some Papilionides, though with others there is a wide
hiatus between these organs to the very base. That the fecal outlet perforates the
abdomen-wall at their point of union seems certain. In both Remus and Ralvphron I
have demonstrated the presence of an orifice, leading from the abdominal cavity between
the uncus and the scaphium, and I have passed a fine needle through it-though, from
the extreme minuteness of the parts, and their dry condition, the demonstration was not
quite so satisfactory as I could wish. Still I can find no anal orifice possible anywhere

else than here. . ,

The tissues of the scaphium become attenuated and expanded bebiud, arching around
as a great concave membrane (fig. 15), which limits the genital cavity, and shuts it off
from the abdomen as by a falling curtain. It is strengthened by slender ribs of chitine, ,
which run down vertically from the cheeks of the scaphium and from the groin of the
keel, somewhat like the nervures in the membrane of the wings, and end in insensible

points, near the penis.

Penis M\ my examinations still leave much obscurity on the form and conditions ofl
the intromittent organ. Its usual appearance is that of a semitube, or gutter, of brown
chitine whose upper (normally and structurally the lower or ventral, for the organ is
reversed) side is widely open, which projects from the curtain outward to near the tip
of the scaphium-keel. But these gaping sides approach each other in various degrees in j
various individuals, the organ at times being nearly flat, especially towards the extremity.
It is then seen to have angular enlargements at intervals, which, when the sides approxi-
mate appear as sharp projecting ribs. The tip is bevelled off to an acute or subacute point,
Lm Ih side of which a ridge passes down into the hollow (fig. 20). [These characters
were precisely repeated in examples of O. Uahj.hron^

But I have seen specimens in which the trough was more than filled with a substance :
which though it was not white, 1 suspect to have been homologous with that pulpy
matter which I have repeatedly described-the more, since in this case too a portion of:
it was extruded to some distance from the extremity.

It is difficult to imagine how an open trough can be efficient for the conveyance oft
the semen into the S vagina. But I venture to suggest though I have no proof, that
in life the tube may be completed by an excessively subtle film, which, in drying, falls
on the opposite chitine-surface, so as to be no longer distinguishable.

IN CERTAIN LEPIDOPTEliA. 337

The falling curtain of the scaphium-base forms below its middle an enveloping sheath
for the penis, which sheath then runs back around it to the very origin of the organ.
This sheath is close, yet wholly free, allowing independent action to the penis, while
it is attached on each side, by a slender ligament, to the floor of the cavity.

Tracing back the organ to its origin, we see that it is fastened by muscles to a thin but
wide ridge of chitine, that springs from the bottom of the seventh segment, and stands up
erect transversely, yet considerably hollow, the concavity backward. This, I presume,
gives the point of resistance for the retraction of the organ after coition. The penis,
behind this attachment (/. e. thoraxwards) enlarges into a thick and long fleshy bulb,
which seems free in the hollow of the sixth segment, and bends back upon itself.

The eighth segment at its bottom gives rise to a hollow shell-like plate of firm polished
chitine, much larger than that of the seventh just mentioned. This and its fellow form
the floor of the genital cavity, surrounding and closing the bases of the valves. To the
upper edge of this plate the penis is affixed by a muscular or tendinous cord stretched in
the contrary direction to that of the seventh segment. I conjecture that this projects
the organ in coition, as the former retracts it, and that both limit its action to the
median line of the body.

Hinge-knobs. Within the plate last described, nearly close to it, and nearly parallel
with it, but quite free from it, is that curious piece, on each side, which I have called a
knob, but which (here, in Remus, certainly) is hollow — a very firm and stout chitinous
shell, to which the vcntro-basal edge of the valve is closely articulated, and from which
the root of the harpe springs.

The small italic letters denote : — a, the uncus ; b, the chitinous lamina of the sca-
phium ; c, its cheek ; d, the sulcus ; e, the aristae ; /, the pentagon ; g, the keel ; h, its
lobes ; i, its groin.

I add some results from late examinations of a few species of Papilio, as chiefly
shedding light on the structure and relations of the scajfliiutn.

P. Ascalaphus. The uncus, scaphium, curtain, and penis, can all be resolved into the
Remus type. The cheek, instead of being erect, leans horizontally outward, making the
sulcus much broader ; the front edge of the cheek is diminished to a slender acute tooth
of black chitine; and its hinder portion (= the sulcus side in Remus) is here the semi-
globose " boat-gunwale " of brown chitinous membrane, or thin horny plate, so corru-
gated as to make strong oblique ridges, sloping inward and forward, on each side of the
shallow sulcus. The keel, though somewhat changed in form, seems homologous ; but
the pentagon appears quite wanting.

In P. Mayo, and in P. Memnon, the general structure is the same ; the differences of
detail are slight and unimportant.

P. Agamemnon. The back of the upper part of the scaphium is elevated, skull-like,
and is thickened into a ridge, where the under surface of the imcus-base is attached to it,
I think, organically, as may be seen when the uncus is forced off. This ridge descends
obliquely on each side, and forms the lateral margin of the curtain, reaching below the

SECOND SERIES. — ZOOLOGY, VOL. II. 47

338 ME. P. H. GOSSE ON THE CLASPING-ORGANS

penis. From the lower edge of each cheek descends a prominent high ridge, which (and
its fellow), the intervening menihrane heing arched, embrace without contact the penis,
sending forward on each side that singularly elegant heart-shaped bundle of white flossy
filaments which I have described above (see p. 311 supra, and PI. XXXI. fig. 7). These
heart-shaped bundles unite into one beneath the penis, and merge into a horizontal
floor of the sheath, which runs back into the cavity of the abdomen. Another less
delicate bunch of similar floss is given off on each side from the outer part of the curtain.

P. Anchisiades. The scaphium-sides go back into the ridge of the curtain (as in Agca
memnon), and sheathe the penis. The scaphium-keel is slit vertically at the front, the
slit expanding upwards. It is equivalent to the same part in 0. Remus; but there is no
pentagon, it scorns quite aborted. The cheeks are rounded, polished, light brown, and
thus chitinous, with a tuft of very short and line aristae, seated on the inner curve, to be
detected only by very careful focusing against the light, but then indubitably.

P. Maccdo,i. Here, again, as I find by several examples, the scaphium is much the same
as in O. Remus, only that the front portion of the check is hardened and pointed, and so
excised as to form the principal tooth. Behind this is a ridge, which runs in an inner
line; the aristae are set from the interior side of tin; tooth, all along the edge of this
secondary ridge. Viewed in front, the appearance much resembles that of 0. Remus;
but the pentagon seems reduced to a mere line, not (visibly) split. The three teeth
which I had described and figured, at the point of the united cheeks (see p. 319 supra,
and PI. XXXII. fig. 3), now no longer appeared; their semblance was probably illu-
sory. The descent to the curtain is normal.

These repeated observations make it highly probable that throughout the genus
Papilio, as in Ormthoptera, the scaphium has consimilar relations with the surrounding
organs, that it is constructed on a common plan though subject to many variations in
details, and that it is throughout composed of three distinct portions, which probably
possess distinct functions.

IX CERTAIN LEPIDOPTEEA. 339

DESCRIPTION OF THE PLATES.

[The figures of the valves have been drawn to about six diameters each, and the others
enlarged proportionally but not to scale.]

Plate XXVI.

Fig. 1. Ornithoptera Armaria. — The interior of the right valve with its harpe in situ, magnified 6 times.

Fig. 2. 0. Armaria. — The harpe, separated and viewed as a transparent object, further magnified.

Fig. 3. 0. Armaria. — The whole genital apparatus, seen from the right side, the right valve having been
removed.

Fig. 4. Ornithoptera Pronomus. — The disk-like extremity of the right harpe, viewed transparently, mag-
nified about 40 diameters.

Fig. 5. Ornithoptera Richmondia. — The disk of the right harpe, magnified about 40 diaui.

Fig. 6. Ornithoptera Haliphron. — The interior of the right valve with its harpe.

Fig. 7. 0. Haliphron. — The ultimate segments of the abdomen, and the contained organs, viewed from
the right side: the right valve removed; the left with its harpe, in situ, represented in faint
outline. The uncus, scaphium, and penis; the last surmounted by the unknown white
substance.

Fig. 8. Ornithoptera Darsius. — The interior of the right valve, with its harpe.

Fig. 9. 0. Darsius. — The harpe, viewed in the same aspect, further magnified.

Fig. 10. 0. Darsius. — The same, viewed sidewise.

Fig. 11. O. Darsius.— The extremity of the penis viewed nearly in front, and with most of the white
matter removed.

Fig. 12. Ornithoptera Rhadarnanthus. — The scaphium and uncus seen from above.

Fig. 13. O. Rhadarnanthus. — The interior of the right valve with its harpe.

Fig. 11. 0. Rhadarnanthus. — The right harpe further magnified.

Fig. 15. 0. Rhadarnanthus. — Teeth of harpe, viewed laterally, and more highly magnified.

Fig. 16. 0. Rhadarnanthus. — The scaphium and uncus, together with the penis, seen from the right side.

Plate XXVII.

Fig. 1. Ornithoptera Heliacon. — The interior of the right valve, with its harpe.

Fig. 2. 0. Heliacon. — Disk of same harpe, further magnified.

Fig. 3. Ornithoptera Heliaconoides. — The interior of right valve, with its harpe.

Fig. 4. 0. Heliaconoides. — The harpe, further magnified. The outlines around the central figure repre-
sent the forms of the teeth opposite to them when these are viewed to advantage.

Fig. 5. Ornithoptera Brookeana. — The interior of right valve, with its harpe.

Fig. C>. O. Brookeana. — The marginal teeth of the latter, more highly magnified, and viewed nearly in the
plane of the harpe, so as to show them in elevation.

Fig. 7. 0. Brookeana. — The scaphium, viewed vertically from above.

Fig. 8. O. Brookeana. — The scaphium, seen from the right side. The left valve ami its harpe are indi-
cated in faint outline.

Fig. 9. Ornithoptera Amphrysus. — The interior of the right valve with is harpe, magnified about 7 times.

Fig. 10. 0. Amphrysus. — The abdominal organs, seen from above.

Fig. 11. — O. Amphrysus. — The abdominal organs, seen from the sides, both valves having been removed.

Fig. 12. Ornithoptera Remus. — The right valve and harpe.

Fig. 13. O. Remus. — The harpe, more magnified.

Fig. 14. 0. Remus. — The same, in outline, flattened.

17*

340 ME. P. H. GOSSE ON THE CLASPING-OKGANS

Fig. 15. 0. Remus.— The end of the abdomen, both valves removed, arid the walls of the segments in

part broken away, to show the scapliiura-curtain.
Fig. Kb 0. Remus. — The uncus and scaphium, from above.

In figs. 16, 17, 18, and 19 the small letters have tbe following signification :— a, the uncus;
b, the chitinous lamina of the scaphium ; c, its cheek ; d, the sulcus ; e, the aristae ; /, the pentagon j
g, the keel; h, its lobes; i, its groin.
Fig. 17. 0. Remus. — The same, from the right and somewhat from above.
Fig. 18. 0. Remus. — The scaphium, from the front.
Fig. 19. 0. Remus. — The scaphium and its keel, from beneath.
Fig. 20. 0. Remus. — The penis, extracted.

Plate XXVIII.

Fig. 1. Papilio Memnon.— The interior of the right valve, with its harpe in situ.

Fig. 2. P. Memnon.— The terminal portion of harpe, shown laterally, more magnified, and viewed as a

transparent object on the stage of the microscope.
Fig. 3. P. Memnon.— The abdominal organs. The right valve having been detached, the left and its |

harpe indicated in outline.
Fig. 4. Papilio Mayo. — The interior of the right valve, with its harpe in situ.
Fig. 5. P. Mayo. — The harpe detached, and more magnified.
Fig. 6. P. Mayo.— The abdominal organs in situ, from the right, the valves having been removed., and

the segments denuded of hair-scales.
Fig. 7. P. Mayo. — The scaphium, more magnified, seen from the front.
Fig. 8. /'. Mayo. — The scaphium, seen from the right.
Fig. 9. Papilio Erechtheus.— The interior of the right valve and its harpe.
Fig. 10. P. Erechtheus— The harpe, more magnified, and viewed from the ventral side.
Fig. 11. P. Erechtheus.— The uncus and its abdominal origin, denuded of scales, viewed vertically from

above.
Fig. 12. P. Erechtheus. — The same, seen laterally from the right.

Fig. 13. P. Erechtheus.— -The organs in situ, with the right valve removed, the left with its harpe, deli-
neated in faint outline.
Fig. 14. Papilio Rhetenor. — The interior of the right valve, with its harpe in situ.
Fig. 15. P. Rhetenor.— Expanded extremity of harpe, further magnified.
Fig. 16. P. Rhetenor. — The abdominal organs.

Fig. 17. Papilio Deiphontes. — Tbe interior of the right valve, with the harpe in situ.

Fig. 18. P. Deiphontes. — The harpe removed from its attachment, viewed from the ventral side, and
further magnified, the terminal portion alone being depicted.

Fie,-. 19. P. Deiphontes.— The uncus, scaphium, and penis in situ, viewed from the right side, both valves
being removed.

Fig. 20. Papilio Schmeltzi. — Interior of right valve, with the harpe in situ.

Fig. 21. P. Schmeltzi. — Interior of left valve, with the harpe.

Fig. 22. P. Schmeltzi.— The extremity of the right harpe, showing the teeth, magnified 140 diameters.

Fig. 23. Papilio Polymuestor.— The interior of the right valve, with the harpe in situ.

Fi°". 24. P. Polymnestor. — The harpe further magnified.

Fig. 25. Papilio Protenor. — The interior of the right valve, with the harpe in situ.

Fig. 26. P. Protenor. — The harpe further magnified.

Fig. 27. Papilio Ascalaphus. — The interior of the right valve, with the harpe.

Fig. 28. P. Ascalaphus.' — The harpe further magnified.

IX CERTAIN LEPIDOPTEEA. 341

Plate XXIX.

Fig. 1. Papilio Helenus. — The interior of tlic right valve, with the harpe in situ.

Fig. 2. P. Helenus. — The harpe, viewed from the side.

Fig. 3. P. Helenus. — Abdominal organs, the right valve being removed.

Fig. 4. Papilio Menestheus. — The right valve, with the harpe in situ.

Fig. 5. P. Menestheus. — The harpe, further magnified.

Fig. C>. P. Menestheus. — The abdominal organs.

Fig. 7. Papilio Pammon. — The interior of the right valve and its harpe.

Fig. 8. P. Pammon. — The harpe, further magnified.

Fig. 9. P. Pammon. — The abdominal organs ; the segments denuded of clothing-scales, and both valves

removed.
Fig. 10. Papilio Agavus. — The interior of the right valve, with the harpe and its affixing base in situ.
Fig. 11. P. Agavus.— The left valve. (See p. 30.2, note.)
Fig. 12. P. Agavus. — The harpe of the same removed, and further magnified.
Fig. 13. P. Agar ns. — The abdominal organs.

Fig. 14. Papilio Machaon. — The interior of the right valve, with the harpe in situ.
Fig. 15. P. Machaon. — The harpe removed and further magnified.

Fig. 16. P. Machaon. — The abdominal organs, viewed from the right side; both valves removed.
Fig. 17. Papilio Arcturus. — The interior of the right valve with its harpe.
Fig. 18. P. Arcturus. — The harpe, further magnified.

Fig. 19. P. Arcturus. — The teeth of the same, with the grooves, still more magnified.
Fig. 20. Papilio Merope. — The interior of the right valve, with its harpe.
Fig. 21. P. Merope. — The harpe, more magnified.

Fig. 22. P. Merope. — The uncus (denuded) and the scaphium, viewed vertically.
Fig. 23. P. Merope. — The abdominal organs, viewed from the right side. The edge of the tegumen

denuded, the right valve removed.
Fig. 21. Papilio Brutus. — The interior of the right valve, with the harpe.
Fig. 25. P. Brutus. — The harpe, more magnified.

Plate XXX.

Fig. 1. Papilio Homerus. — The interior of the right valve, with its harpe.

Fig. 2. P. Homerus. — The harpe removed and further magnified.

Fig. 3. P. Humerus. — The abdominal organs, viewed from the right side ; both valves removed and

the edges of the segments in part denuded.
Fig. 1. P. Homerus. — The scaphial teeth, seen from the right, and further magnified ; the left double

tooth in faint outline only.
Fig. 5. Papilio Ulysses. — The interior of the right valve, with the harpe.

Fig. G. P. Ulysses. — The extremity of the harpe, viewed at a different angle, and more magnified.
Fig. 7. P. Ulysses. — The abdominal organs, viewed from the left side, partly denuded, and both valves

removed.
Fig. 8. P. Ulysses. — The point of the nearer scaphial tooth, in outline, much magnified.
Fig. 9. Papilio Phorcas. — The interior of the right valve, with the harpe.
Fig 10. Papilio Thoas. — The interior of the right valve, with the harpe.
Fig. 11. P. Thoas. — The abdominal organs: the ultimate segment denuded of scales, and both the

valves removed.

342 ME. P. H. GOSSE ON THE CLASPING-OEGAXS

Fig. 12. Papilio Turnus. — The interior of the right valve, with the harpe.

Fig. 13. P. Turnus. — The ventral edge of the harpe, with its spine, viewed from the dorsal side, almost

horizontally, and a little more magnified.
Fig. 14. P. Turnus. — The dorsal edge of the harpe, with its compound spine, viewed from the ventral

side.
Fig. 15. P. Turnus. — The abdominal organs in situ, both valves having been removed.
Fig. 10. Papilio Polycaon. — The interior of the right valve, with the harpe.
Figs. 17 & 18. P. Polycaon.' — The spinous knob, under its two prevalent forms, more magnified.
Fig. 19. P. Polycaon. — The abdominal organs in situ, the right valve having been removed.
Fig. 20. Papilio Axion. — The interior of the right valve, with the harpe.
Fig. 21. Papilio Hesperus. — The interior of the right valve, with the harpe.
Fig. 22. P. Hesperus. — The harpe, viewed from the ventral side, and further magnified.
Fig. 23. P. Hesperus. — A portion of a curved ridge of the harpe, showing teeth on an enlarged scale.
Fig. 24. P.Hesperus. — The abdominal organs, viewed from the- right; both valves removed, and the

segmental edges almost denuded.
Fig. 25. Papilio Demoleus. — The interior of the right valve, with the harpe.
Fig. 2G. P. Demoleus. — The ridge of the harpe, with its girding ribs and terminal tooth, much more

magnified.
Fig. 27. /'. Demoleus. — The extremity of the abdomen, with its apparatus; the segments denuded, and

both valves removed.

Plate XXXI.

Fig. 1. Papilio Bathycles. — The interior of the right valve, with the harpe in situ.

Fig. 2. ]\ Bathycles. — The harpe, much more magnified.

big. .'3. /'. Bathycles. — The scaphium, viewed vertically.

Fig. 4. P. Bathycles. — The abdominal organs, viewed laterally.

Fig. 5. P. Bathycles. — The tip of the penis, further magnified.

Fig. G. Papilio Agamemnon. — The interior of the right valve, with the harpe.

Fig. 7. P. Agamemnon. — The abdominal segments (denuded of scales and hairs except on the uncus), the

scaphium, the penis and its sheath.
Fig. 8. P. Agamemnon. — The uncus, vicwred vertically.

big. ',). Papilio Erithonius. — The interior of the right valve, with the harpe.
big. 10. P. Erithonius. — The elevated triple wall of the harpe, viewed from the dorsal side and much

more magnified.
Fig. 11. P. Erithonius. — The abdominal organs, viewed laterally, the surfaces denuded of the copious

scales and hair-scales, and both the valves removed.
Fig. 12. P. Erithonius. — The chitinous curtain described at page 315, supra.
Fig. 13. Papilio Anchisiades . — The interior of the right valve, with the harpe.
Fig. 14. P. Anchisiades. — The harpe removed, and further magnified.
Eig. 15. P. Anchisiades. — The scaphium, viewed from above ; the terminal portion, but not the extreme

point, of the uncus seen overarching the middle.
Fig. 16. Papilio Children^. — The interior of the right valve, with the harpe.

Fig. 17. P. Child Irenes. — The terminal hand of the harpe in another aspect, much more magnified.
Fig. 18. P. Childrence.- — The abdominal apparatus, the right valve removed, the left in situ, outlined.
Fig. 19. Papilio Sesostris. — Outline of the right harpe, showing the teeth.
Fig. 20. Papilio Vertumnus. — The interior of the right valve, with the harpe.

IN CERTAIN LEPIDOPTERA. 343

Fig. 21. P. Vertumnus. — The abdominal organs, the right valve removed, the left outlined.

Fig. 22. Papilio Lysander.— -The interior of the right valve, with the harpe.

Fig. 23. P. Lysander. — The extremity of the harpe, much more magnified.

Fig. 24. P. Lysander. — The uncus, with its toothed scaphium, and the penis laterally split and

emptied.
Fig. 25. Papilio Euryleon. — The interior of the right valve, with the harpe.
Fig. 26. P. Euryleon. — The harpe, removed with its muscular basis attached, further magnified.
Fig. 27. Papilio Err/deles. — The interior of the right valve, with the harpe.
Fig. 28. P. Ergeteles. — The terminal half of the harpe, further magnified.

Fig,

6.

Fig.

7.

Fig.

8.

Fig.

9.

Fig.

10

Fig.

11.

Fig.

12

Fig.

13

Plate XXXII.

Fig. 1. Papilio Macedon. — The interior of the right valve, with the harpe.

Fig. 2. P. Macedon. — The harpe detached, and more magnified.

Fig. 3. P. Macedon. — The abdominal organs, both the valves having been removed.

Fig. 4. Papilio Bromius. — The interior of the right valve, with the harpe.

Fig. 5. P. Bromius. — The harpe, removed, seen laterally from the ventral side of the valve, and more

magnified.
P. Bromius. — The abdominal organs, with the left valve /'// situ, seen from the right side.
Papilio Podidirins. — The interior of the right valve, with its harpe.
P. Podalirius. — The harpe, further magnified.
P. Podalirius. — The tegumen and scaphium, viewed vertically from above.

P. Podalirius. — The same with the penis, viewed laterally from the right ; the valves removed
and the segments denuded of scales in both figures.

Papilio Latreillii. — The interior of the right valve, with the harpe.

Papilio Seiuperi. — The interior of the right valve, with the harpe.

P. Semper/. — The harpe, removed, and further magnified.
Fig. 14. P. Semperi. — The abdominal oi'gans viewed from the right side, both the valves having been

removed.
Fig. 15. Papilio Lycidas. — The interior of the right valve, with its harpe.

Fig. 16. P. Lycidas. — The abdominal organs, the valves removed, and the segments nearly denuded.
Fig. 17. Papilio Doubledayi. — The interior of the right valve, and its harpe.
Fig. 18. P. Doubledayi. — The second aspect of the penis described at page 323.
Fig. 19. Papilio Rhodifer. — The interior of the right valve, with its harpe.
Fig. 20. P. Rhodifer. — The extremity of the penis, further magnified.
Fig. 21. P. Rhodifer. — The abdominal organs, the right valve having been removed, the left shown in

outline.

Papilio Archesilaus. — The interior of the right valve, with its point of hair-scales, and the harpe.

P. Archesilaus . — The harpe detached, viewed obliquely, and further magnified.

P. Archesilaus. — The abdominal organs, the right valve removed, the left shown in outline.

Papilio Zalmoxis. — The interior of the right valve, with the harpe.

P. Zalmoxis. — The tip of the harpe, more magnified.

P. Zalmoxis. — The same, seen in a different aspect.

P. Zalmoxis. — The abdominal organs, from the right ; the right valve removed, the left shown

in outline.

Fig.

22.

Fig.

23.

Fig.

24.

Fig.

25.

Fig.

26.

Fig.

27.

Fig.

28.

344 ME. P. H. GOSSE ON THE CLASPING-ORGANS

Plate XXXIII.

Fig. 1. Papilio Policenes. — The interior of the right valve, with its harpe.

Fig. 2. Papilio Ucalegon. — The interior of the valve, with its harpe.

Fig. 3. P. Ucalegon. — A lateral view of the harpe, separated from the valve ami viewed dorsally.

Fig. 4. Papilio Agesilaus. — The interior of the right valve and its harpe.

Fig. 5. P. Agesilaus. — The abdominal organs, viewed laterally.

Fig. (3. r. Agesilaus. — The uncus, viewed vertically.

Fig. 7. Papilio Parmatus. — The interior of the right valve, with its harpe.

Fig. 8. Papilio Codrus. — The interior of the right valve, with its harpe.

Fig. 9. P. Codrus. — The harpe, further magnified.

Fig. 10. P. Codrus. — The termination of the abdomen.

Fig. 11. P. Codrus. — The penis, much magnified.

Fig. 12. P. Codrus. — Another view of the same, much magnified.

Fig. 13. Papilio Nireus. — The interior of the right valve, with its harpe.

Fig. 14. P. Nireus. — The harpe removed, seen obliquely sidewise, somewhat more magnified.

Fig. 15. P. Nireus. — The spathulate tip of the harpe, seen from above, still more magnified.

Fig. 10. P. Nireus. — The uncus, the scaphium with its processes, and the penis, seen laterally ; the

segments partly denuded, and both valves removed.
Fig. 17. Papilio Diphilus. — The valueless harpe (?), magnified .'20 times.
Kg. 18. P. Diphilus. — The same, on the scale generally adopted.
Fig. 19. P. Diphilus. — The spinous knob, more magnified.
Fig. 20. P. Diphilus. — The uncus, scaphium, ami penis in situ; the thatch of hair iu part removed, and

also the right harpe. The knob of the left harpe is seen in contact with the notched truncate

end of the uncus.
Fig. 21. Papilio Pohjdorus. — The abdominal organs, denuded of scale-hairs.
Fig. 22. P. Polydorus. — The right scaphial tooth (?), seen from within.
Fig. 23. P. Pohjdorus. — The left valve, seen from within.
Fig. 24. Papilio Anterior. — The interior of the right valve, with the harpe.
Fig. 25. P. Antenor. — The uncus, with its scaphium, seen from above.
Fig. 26. P. Antenor. — The uncus, with its scaphium, seen from the right side.
Fig. 27. Papilio Hector. — The parts in situ, seen from the right side, when the thatching of hair is

removed.
Fig. 28. P. Hector. — The uncus, as seen from behind.
Fig. 29. P. Hector. — The right scaphial tooth (?), seen from below.
Fig. 30. P. Hector. — A valve, detached.
Fig. 31. P. Hector.— The penis.

IN CERTAIN LEPIDOPTERA.

345

INDEX.

\ < ontropus, analogy of, 270.
Anthea, tegumen of, 332.
Anus, site of, 276, 296, 304
Asymmetry, accidental, 303; in Hespe-

riadce (ftnote) 298; in valves, 298.
and front," terms how used, 273,

296,305,324,337.
Burmeister, on general turgescence, 281 '>:

on harpe, 273; on meconium, 269,

288 : on valve, 2G5.
Calliili . 332.

Colias, pn
Curtain, in 0. /. : in P. Eri-

273 ; on P. Puli/-
doru . 6 ; on

ium, 277.
r,
Eaton, on Spermatozoa in Eph mera,2&8.
Gonepteryx, pn In nsors, 3! '<■
271,313,31

72,31o,31i
319, 7; : of, 272,

I, 2 3, 319, 325, 32 . 337; cha-
racter of, 271 :

299, 331 : type-forms

Heboi i i 000.

Herold, on /'

III sp< : ;

Leon- 1 --7.

Man: ;
Mecoi ium, :

288, 311.
Morpho, pi 334.

Ornithi

293; An '

of, 285; Helia on, -I'D; Hi li

21 0 , Pi :' : Remu: ,

hilt ■

tliu-. -

Ricl 34.

Papilio Agamemnon, 314, 307 ; Agavus,
301; Ag silaus, 326; Anchisiades,
315, 337; Androgeos, 310; Antenor,
331; Arcb silaus, 324; Arcturu
Ascalaphus, 300, 337 ; Axion, 311 ;
Bathycles, 313; Bromius, 268, 320;
Brutus, 306; Chaon,300; Childrenae,
316; Codru . 327; Coon 267; Dei-
phontes, 297; Demi 312;

Diphilus, 329, aberi . 329 ;

Doubledayi, 323; I, , 296 ;

Ergetele ,319; Erithonius, 268, 314;
Euryleon, 318; Gigon, s]

!1; Helenus,
30.1: Hesperus, 312; Hon
Latreillii, 321 ; Lycophron, 310; Ly-
sander, .'117: Macedon, 319, 338 ;
. 295 ; Mi ■union,
294, continental forms of, 295; Me-
nestheus, 301 ;
lus, 300; Nin as,

310; Pamm . 301 ; Pi matu , 327;
Phoi i . Podaliriu

Policeues, 325 ; Polycaon, 310 ;
dorus, 329; Polymnestor, 299; Pro-
tenor, 299; Rhetenor, 297 ;
fer, 020 ; Sc!i

322; Sesostris, 316; Thersite .
Tlioas, 308; Turnus, 309; I
325 ; My es, 307 ; Vertumnus, 317;
X ilmoxi . irs in,

21 9.

Penis, i' at, 27 - : ba e, 337 :

changes in, 2 7. (/ 0, 318,

323, 336; characters of, 277,
forms of, 278, 287, 321, 324, 32S

i 322 ; Owen on, _". i ;

position nf. 277: protrusion of, 277,
313,316, 319, 320, 323,324; sheath
of, 2:i7, 313, 314, 3 in O.

Haliphron, 285, 204: in O. Remus,
-■ks, 323; in P.Lysan-
der, 318; in P. Rhodifer, 324; in P.
& mperi, 322.

Pieridee, Herold on, 265.
. 332.

Prehensors, material, 268; use for dia-
gnosis, 268; variety of, 207, 273.

Scales, foreign, 269, 27;!. 2-0, 209, 305,
307, 311,319.

Scaphium, bristles of, 276,285, 2 S9, 297,

310, 316, 320, 325 ; characters of, 275;
.. of, 335; ci lour of, 270, 289,

305, 310, 325, 320, 330; forms of,
275, 313; function of, 276, 305, 306;
in O. Remus, 335 ; in P. Agamemnon,
337; in P. Anchisiades, 338; in P.
Ascalaphus, 337 ; in P. Codrus, 328;
in P. Macedon, 338; in Pieridae, &c.,
332; keel of, 2-::, 289,292,301,309,
335; pentagon, 335 ; rami. 276, 296,
207, 300, 304, 313; resemblance to a
boat, 275, 200, 304, to a jaw, 275,
295, 304; tooth of, 275, 296, 304,
305, 313, 330, 331.

Scudder and Burgess, on Nisoniades,
266; on asymmetry, {ftnote) 298.

Spermatozoa, Gladstone on, 287 ; in
Neuroptera, 288; in penis, 278,285,
286, 288, 316,318.

Terias, prehensors, 333.

Trichoptera, McLachlan on, 277.

. 313, 31 7, 321, 328; cha-
racters of, 273; forms of, 274. 329,
in O. R mu 1, 335; ininiii . 325 ;
re emblances, 274.

Uniformity, individual, 1 ftnote) 289,
020.

Urine, in Insects, 287; in Serpents,
288.

Valves, asymmetry in. 298; characters
of, 270; disguises of, 329, 330, 331 ;
hinge-knobs, 337; Wallace on, 271.

Variety in structure, 267, 270: object
of, 279; specific, (ftnote) 289, 300,

311, 017.

Wallace, on Ornithoptera, 271 ; on P.

Memnon, 295.
White, on Rhopalocera, 265, 267; on

P. Podalirius, 32 1 .

SECOND SERIES. — BOTANY, VOL. II.

48

[ 347 J

X. On certain Points in the Anatomy of the Polynoina, and on the Polynoe' (Lepidonotus,
Leach) clava of Montagu. By Alfred Gibbs Bourne, B.Sc. Lond., Univ. Scholar
in Zoology, and Assist, in the Zoological Laboratory, University College, London.
(Communicated by Prof. E. Ray Lankester, M.A., F.L.S.)

(Plates XXXIV.-XXXVI.)

Bead January 18th, 1883.

THE investigation of the anatomy of the Polynoe clava of Montagu was undertaken, at
the suggestion of Professor Lankester, by Mr. W. B. Benham, Demonstrator in Zoology,
University College, London. Mr. Benham made a large series of drawings, comparing
this species with Polynoe squamata ; but while working at the Zoological Laboratory at
Naples, I was enabled to examine a very large number of living specimens of Polynoe,
more especially of P. grubiana, Clap., which I shall show to be only a variety of our
P. clava ; and as I had obtained certain new results with regard to the anatomy of that
and other species of Polynoe, Mr. Benham kindly placed his notes and drawings in my
hands.

I do not propose to deal at any length with questions of synonymy. The enormous
mass of literature upon the group, the number of species which are now known to exist,
and the minute characters which separate them, combined with the very general cha-
racter of most of the descriptions of the older writers, render it undesirable to do so on
the present occasion. Polynoe clava was first described from the coast of Devonshire,
by Montagu * as Aphrodita clava ; and he fully distinguished between this form and
P. squamata, Linn. ; his single figure, however, is insufficient.

Johnston t also distinguishes between this form and P. squamata, and gives the fol-
lowing characters : —

" L. squamatus. — Scales ovate and reniform, imbricate, granulous, ciliated on the outer margin ;
tentacula and tentacular cirri incrassated below the point ; bristles of the ventral branch with sharp
denticles on the thickened portion of the 'shaft below the smooth talon-like apex. Length 1-2";
breadth 3'"."

" L. clava. — Scales subcircular, margined, not imbricate nor meeting on the mesial line, maculated,
unciliated on the edge ; tentacular and dorsal cirri bulbous below the point ; bristles as in L. squamatus.
Length IV; breadth 4"'."

Claparede seems to have overlooked the species entirely, and describes his P. grubiana
as differing from P. squamata, Linn., in its elytra not covering the back, in the smooth
contour to the elytra, and in the papillate character of the palps.

Mcintosh \ points out that Claparede's P. grubiana may be a variety of Lepidonotus

* Montagu, Linn. Trans, ix. p. 108, t. 7. fig- 3.
t Johnston, ' A Catalogue of British Non-parasitical Worms,' 1S65.
X ilclntosh, Trans. Zool. Soc. is. 1877, p. 371.
SECOND SERIES, ZOOLOGY.— VOL. II. 49

348 ME. A. G. BOUENE ON CEETAIN POINTS

clavus, " since lie does not distinguish it from the latter while contrasting it with
L. squamatus." P. squamata, as Malmgren has pointed out, is not found in the Mediter-
ranean ; while, at any rate at Naples, P. grubiana is the commonest of the Polynoina.
I have collected there about a hundred specimens ; and thus, while I am convinced that
it is the same species as our P. clava, I have been startled to observe the great individual
variation which may occur in a single species.

The Mediterranean forms possess rather darker scales, but possess in other places less
pigment than do the specimens we had obtained from Devonshire ; they are moreover
considerably smaller.

My observations upon the series of P. clava have led me to classify their characters
thus : —

Variable.

Size and shape of the body.

Elytra, in respect of their size, shape, and tint.

Amount of medio-dorsal region of back covered
by them ; extent to which each overlaps the
following elytron ; firmness of their attach-
ment to the body.

Thickness of the anterior margin of the pro-
stomium, influencing the apparent place of
origin of the lateral tentacles.

Amount of pigment on the body-surface.

Presence of acicula in the pygidial segment.

Constant.

Number of segments in the body.

Number of elytron-bearing segments.

Structure of median tentacle.

Structure of lateral tentacles.

Structure of palps.

Shape of equivalent parapodia.

Structure of the notopodial and neuropodial cirri.

Structure of the ventral tubercles (nephridial
papillae) .

Character of equivalent setse.

Fringing of the margin of the elytron (not in-
cluding the tactile papilla?) .

Position and size of the eyes.

Mode of origin of the median prostomial tentacle.

Position of the anus.

Structure of pygidial cirri.

The following variations are also to be found in one and the same individual : —

Structure of the notopodial and neuropodial setse ; their length and number.
The elytra in respect of size, shape, and amount of granulation on their surface.
The shape of the parapodia.
This individual variation necessitates great care, in making comparisons between these
forms, that exactly ecurivalent structures be compared.

Polynoe clava, Montagu.

Shape of the Body. — As I have pointed out, we can attach very little importance to such
difference in the shape of the body as Johnston has described as existing between P. squa-
mata and P. clava ; my specimens of P. clava show much more variation among themselves
(compare figs. 3 & 4, PI. XXXIV.). Still less can any importance be attached to the
position of the elytra with regard to the medio-dorsal region ; whether these latter do
not meet, simply meet, or overlap, depends entirely upon the amount of food in the
alimentary canal and the number of ova contained in the body-cavity. Figs. 3 & 4 show
two extremes in this respect.

The body consists of a prostomial lobe with its tentacles and 27 somites ; the 1st of these
is the peristomial somite (Huxley), the 2nd the first elytron-bearing somite, and the
27th the pygidial somite. All these segments bear more or less modified parapodia.

IN THE ANATOMY OF THE POLTNOINA. 349

Parapodia. — The parapodia in the central region are normal, but at either end of the
body they become modified. The normal parapodium consists of notopodiiun and neuro-
podium, the latter being about four times the size of the former, which, however, in
P. clava cannot be regarded as a mere tubercle projecting from the upper surface of the
neuropodium (Huxley, P. squamata), but more nearly retains its normal position as a
lobe of the parapodium corresponding to the neuropodial lobe.

The notopodial acicula is well developed. The notopodial seta? are about twelve in
number ; but this number is by no means constant, as they are continually falling out and
being replaced by new ones ; consequently setae of various stages of growth may be
found in one parapodium ; there appear, moreover, to be permanent varieties in the
notopodial series. Figs. 12, 12 a, 12 b, & 12 c represent such varieties of the notopodial
seta? of P. squamata (only the extremities are drawn). Figs. 12 & 12 a come from
short seta3, figs. 12 b & 12 c from very long ones. In the short seta? there are about
25 pectinated frills, which encircle in the central region about two thirds of the cir-
cumference of the blade ; but the upper frills, which are more pectinated, encircle much
less, and the lower ones rather more ; there is only one longitudinal series of frills.
The long seta? present series of frills which are not so well developed. The notopodial
seta? of P. clava are all formed upon the type shown in fig. 11, corresponding to the
long seta? of P. squamata.

The neuropodial setae are very similar in two species (PI. XXXV. figs. 13 & 14). There
are two longitudinal rows of 7 or 8 pectinated frills placed opposite to one another^
although the fact that the ends of the frills running slightly downwards may give in some
series the appearance of the frills alternating on the two sides. The distal rows of frills
are more strongly developed than the rest, and present five large teeth upon their free
edge. In P. squamata the central tooth is relatively much larger than in P. clava. The
proximal frills are always feebly developed.

The neuropodial acicula pierces a small papilla; dorsad of this papilla there are
about six neuropodial seta?, and ventrad about twice that number. This, as Grube
points out, is the usual arrangement — twice as many seta? ventrad as dorsad of this
papilla.

The notopodial cirrus is more bulbous below the point in P. clava than in P. squamata.

The neuropodial cirrus of P. clava differs from that of P. squamata in the same way
that the notopodial cirrus does, and is besides shorter and rather stouter.

The papilla (pyriform tubercle of Huxley, ventral papilla of Grube and Ehlers, figured
occasionally without description by Claparede and Malmgren) which I propose to
call the uephridial papilla, springs, as Huxley pointed out, not from the parapodium, but
from the neural surface of the somite itself.

Elytra.— -The elytra present various shapes, according to their position, in both species.
The shape also varies in individuals, as will be seen by comparing figs. 3 & 4, PI. XXXIV.

The anterior pair vary in shape between an oblong with rounded corners (fig. 7) and a
broad ovate (fig. 4).

The three anterior pairs present a slight roughness on the surface ; but the posterior
pairs are almost smooth.

49*

350 ME, A. G. BOURNE ON CEETAIN POINTS

The margin is entire, and not fringed.

In P. squamata (fig. 10) the surface in all the elytra is very rough, presenting large
granular conical lumps, especially developed in the outer and posterior regions of the
scale. Some of these granulations are red in colour, giving the rich grey colour which
the elytra possess in this species. The outer and posterior margins, and in the anterior
pairs the anterior margin also, are richly fringed with long processes, in addition to the
tactile papilke which are scattered over the surface.

Prostomium. — The prostomium differs in shape in the two species (PL XXXVI. figs. 15
& 1G). In P. clava it is much more truncated anteriorly than in P. squamata. In the
latter species the prostomial tentacles are obviously mere prolongations from its anterior
edge. In P. clava the anterior portion is much smaller, and there is a tendency to form
a transverse ridge, giving the impression, from a dorsal view, that the lateral tentacles
spring from underneath the anterior edge of the prostomium. This they undoubtedly do
in many species of Polynoina ; but, from observations which I have made upon unde-
termined species collected at Naples and from the Norwegian Fjords, I have come to the
conclusion that there may be every transitional condition between these two extremes.

Series of transverse sections have shown that these three tentacles are supplied with
nerves directly from the supraoesophageal ganglions, i. e. no doubt have an archi-
cerebral nerve-supply, and so are not of the nature of appendages, although they simulate
them in structure, but are mere prolongations of the prostomium ; they thus have funda-
mentally the same origin in the whole group. The anterior edge of the prostomium may
thin out into the bases of the tentacles ; it may be thicker, and tend to form a transverse
ridge, which may be produced and overhang the bases of the tentacles ; small points may
be formed here ; and, lastly, these may be carried far forward, and so completely hide the
origin of the lateral tentacles. If every intermediate condition between the two ex-
tremes can be found, these can hardly be used in generic distinctions.

In P. squamata the anterior pair of eyes is much larger than the posterior pair, and lie,
as shown in PL XXXVI. fig. 15, very near them ; but in P. clava, while the posterior
pair have the same position, the anterior pair, which are of the same size, lie further
forward.

First Body-Somite, Peristomial Somite (Huxley). — The most anterior somite of the body
possesses much-modified parapodia. A notopodial and a neuropodial lobe may, however,
be distinguished ; and each bears four setae. Huxley has described the existence of
similar structures in P. squamata (an acicula and two minute setae), and points out that
this fact, which Max Miiller * described in his account of the development of Polynoel
proves the peristomial cirri to belong to a modified somite.

Second Body-Somite, or Buccal Somite.- — The second somite is also modified. The noto-
podial and neuropodial lobes are obscurely divided ; both aciculae are present ; the noto-
podial setae have the same general structure as in other somites, but the neuropodial setae
resemble the notopodial in structure rather than the neuropodial of other somites.
This fact, which appears to have been overlooked, is interesting as connecting the normal
parapodium with that of the much modified most anterior body-somite.

* JIiiller's Archiv, 1851.

EN" THE A>"ATO:MT OF THE POLENOENA.. 351

This second body -somite appears to be the first elytron-bearing somite in all Polynoina.

The nenropodial cirrus is much modified, resembling a notopodial cirrus in size
and shape, is bent inwards towards the mouth tag. 1, bue. cr.\ and is the buccal cirrus
XLaberg.). There can be no doubt that this buccal cirrus is a modified neuropodial
cirrus ; and the possibility of a neuropodial cirrus simulating a notopodial (we have seen
a similar ease in the prostomial tentacles simulating notopodial cirri ) is of irreat import-
ance to us in considering the homology of the pygvdial cirri.

Palps. — The palps ditfer from all the other tentacular struct;. 3 being muscular
along their whole length. They are capable of great elongation and contraction.

In P.sgvamata they are smooth: but in P. ; they have six longitudinal rows of

tactile papilla?, each provided with its sensory hair.

These palps present more difficulty than do any of the other tentacles or cirri. They
originate (sections of P. (Sarmothoe) areolata) just where the prostomium joins the
peristomial and buccal somites, although they appear to have more connexion with :
prostomium than with the other somites. Their nerve-supply appears to come from the
snpracesophageal iranslion. Their prostomial connexions are borne out when we com-
pare them with similar organs in such simple forms as the SyUidae, where they are
evidently prostomial, sprii - ° . . ie, loc. cit.) from its ventral surface.

This view is completely borne out by Max Muller's * account of their development.

Pygydial Somite. — The terminal somite is very much reduced, and appears merely -
support for the pygidial cirri. These resemble notopodial cirri in size and structure.
They are well developed in both species under consideration : but in nianv species of
Polynoina they are quite small. In such forms the anus is terminal ; but where they are
so strongly developed the anus is pushed forwards, and comes to lie on the dorsal
surface.

I have found in two or three specimens of P. clava an acicula (fig. 6) running
dorsally, and perforating a well-marked papilla, as do the other acicula?. In other speci-
mens, of which I have examined a large number with this special point in view, there is
a trace of such a structure ; but in the majority all trace is gone; possibly the acicula
drops out while the animal is thus quite young.

In other undetermined species of Polynoe which I have examined there is. as a rule,
a small acicula in some, a small acicula accompanied by rudimentary setae in others,
in the pygidial segment, most strongly marked where the segment is little modified.
The presence of these structures seems to me of great interest, as showing that the
pygidial cirri are the neuropodial (as they are ventrad of the acicula, whether that be
neuro- or notopodial) cirri of a much-modified somite, a pygidial somite. Further, just
as we found the second body-somite presented a connecting link between the first.
the peristomial segment, and the other somites, so in the penultimate somite we find
that the seta? are few in number, and not strongly marked in character, connecting the
pygidial somite with those in front of it.

Nephridial Papilla and Nephridia. — Minute and detailed as many of the systematic
descriptions of various species of Polynoina are, it is not a little curious that the exist-

5 Max Miiiler, Miillers Archi , 1S5]

352 MR. A. G. BOURNE ON CERTAIN POINTS

ence even of these ventral papilla? has heen so very generally overlooked. While
Claparede, Mahngren, and Ehlers insert it occasionally in their figures, the only
references to it are made by Huxley and Grube. Claparede, however, does describe the
papiike in Ilermadion fragile, and suggests that it may be the orifice of the segmental
organ. Huxley *, in his account of the anatomy of P. squamata, which was based upon
his own observations, says : — " Springing from the neural surface of the somite, close to
the parapodium, there is a small pyriform tubercle, divided by longitudinal grooves into
about eight segments. This is possibly connected with the reproductive functions."

Grube f recognizes its wide-spread existence : — " At the point where the parapodium
springs from the ventral wall of the body, exists in all Polynoina a very small papilla
(ventral papilla), which, at any rate in pregnant examples, is perforated, and serves for
the passage of the eggs to the exterior. The animals in which they run out to a point
are probably males. I have repeatedly found sticking to the spots slimy masses, which,
on account of their resemblance to those observed in living Heteronercis forms, I take
to be spermatozoa. This papilla is generally present, but, as a rule, absent upon the first
3-0 seta-bearing segments."

I have had an opportunity of examining a very large number of living Polynoina,
representing some sixteen or seventeen different species, and am in a position to corrobo-
rate Grube's statement as to its universal existence. Both Huxley and Grube accurately
describe its position ; it springs from the ventral wall of the body itself, and not from the
parapodium. It varies very much in size in different species: in P. clava it is large
and thick (PI. XXXVI. figs. 17 & 18) ; in P. squamata it is mucli longer and thinner
(fig. 19) ; in P. spinifera, Ehl., it is very slightly developed (fig. 20) ; and in P. areolata,
Grube (fig. 21), it is merely a slight swelling of the body-wall, which is perforated. In
all the other species examined the extent to which it is developed is intermediate to
these conditions.

As Huxley points out, it may exhibit eight longitudinal grooves upon its surface. At
the breeding-season the body-cavity is filled with the generative products, rendering the
body-wall perfectly tense, and they are forced into the space in the wall of the papilla ;
the grooves then disappear. The constancy of their number points to some structural
basis ; but their absence in such a large number of species indicates that they have no very
great significance.

The papilla} are absent, as a rule, from the eight most anterior somites, and are con-
tinued down to the penultimate somite. In P. clava there are eighteen pairs.

Careful and continued examination of various species has proved that they are not
connected with the reproductive functions, but are the apertures of the nephridia or
segmental organs, as Claparede surmised in his Hermadioti fragile ; serial sections
combined with examination in the fresh state, both in the compressorium and after
teazing, have revealed the whole structure of these organs (PI. XXXVI. figs. 22-25).

* Huxley, ' The Anatomy of the Invertebrated Animals,' p. 231.

f Grube, " Bemerkungen iiber die Familie der Aphroditeen, III. Polynoina," Bericht der schlesischcn Gcsellschaft,
187G, pp. 2G-52.

IN THE ANATOMY OF THE POLYNOINA. 353

Ehlcrs * has described a segmental organ in P. pellucida as a contractile sac lying in
the parapodium and opening externally by several (4-6) ciliated mouths upon both its
notopodial and neuropodial faces, the internal opening lying near the dorsal wall of the
somite. Ehlers describes these as occurring in every somite from the second onwards.

Claparede t states that he has observed the ciliated rosettes of Ehlers in P. lunulata,
but, in spite of the transparency of the animal, has been unable to acquire the conviction
of their communication with the segmental organs. I have not studied P. pellucida,
but, from the analogy of other forms, strongly suspect that what Ehlers has seen is not
the segmental organ at all$. The nephridium, in all the forms which I have examined, is a
funnel-shaped tube, the walls of which may be much plaited and folded, but it does not
appear to be at all convoluted. This tube opens internally in the ventral region of the
body, close to the two ventral muscular masses, and then runs posteriorly ventrally out-
wards. Near the base of the papilla its lumen dilates, forming a small vesicle, whence a
straight tube of even diameter traverses the axis of the papilla. There is no muscular
development in connexion with it at any part of its course ; it never enters the para-
podial region at all ; and although ciliated along its whole length from the internal
funnel to the vesicle, no cilia are developed at its external opening.

Fig. 22 is a diagrammatic drawing compiled from the series of sections &c.

Fig. 23 is a transverse section of the papilla, and shows the longitudinal grooves, the
epidermis, which is continuous with the epidermis of the rest of the body, and the ciliated
epithelium cells lining the excretory tube ; between these is a layer of connective tissue'
branched corpuscles in a connective jelly.

Fig. 24 represents a section of the tube near the internal opening, while fig. 25 is an
intermediate section in the region of the vesicle. The body-cavity is packed with ova,
which press the two walls of the vesicle together, and in many cases almost occlude the
lumen. The ova force their way in between the nephridial epithelium and the epidermis,
and thus, on an external examination, give the impression of being within the lumen
of the papilla ; and this appearance it is which has led previous observers to connect these
papilla? with the reproductive function.

How do the generative products pass to the exterior ? There certainly are no special
ducts ; and they cannot either find exit through those nephridia or by pores, temporary
(i.e. ruptures) or permanent, in the body-wall. Ehlers and Sars both mention such
pores as existing in the parapodial wall. The generative products are known to escape
by rupture in other marine Annelids, e. g. Polygordms. All my observations upon living
animals favoured this view. I never actually saw the ova or spermatozoa spontaneously
shed ; but the slightest irritation when the body-wall was tense caused a slight rupture here

* Ehlers, ' Dio Borstenwurmer,' 1864-1863, p. 116.

t Claparede, ' Annelides Chetopodes du Golfe de Naples,' p. 374.

X Since this was written, Mr. "W. A. Haswell, M.A., B.Sc, in " A Monograph of the Australian Aphroditea " (Proe.
Linn. Soc. New South Wales, vol. vii.), has described the segmental organ in P. (Antinoe) prceelara and P. (Antinoe)
Wdhlii allied to P. pellucida, Ehlers. That author has also arrived at the conclusion that Ehlers has not seen the
true segmental organs, but only intestinal cteea ; he describes the former as opening at tho ventral tubercles, but does
not give any figures.

354 ME. A. G. BOURNE ON CERTAIN POINTS

and there, and the products streamed forth. I have kept such animals for a long period
after this, and they have been apparently none the worse. I have certainly never seen
ova or spermatozoa in the lumen of the nephridium.

Fig. 26, representing a longitudinal section of a neuropodial cirrus of P. spinifera, is
inserted to show the large round connective-tissue corpuscles which exist there in addition
to the branched cells. The nerve-filament passing up the axis is also shown.

Following the example of Sars, Claparede, Grube, and Ehlers, I have rejected Kinberg's
and Malmgren's genera, and have usedSavigny's original generic name Polynoe. I have
shown that so many of the characters used by these authors to found generic distinctions
are variable in the same species, and that with regard to others it is easy to find all
intermediate variations between two extreme conditions.

I quite agree with Mcintosh that the utmost importance may be attached to the
exact character of the seta?, but thiuk that he hardly sufficiently insists upon the varia-
tion which occurs in these in individuals, and the necessity of comparing equivalent seta?.
The seta?, for instance, in the neuropodium of the third body-somite are totally unlike
the neuropodial seta? of the somites which follow, and much more resemble notopodial
seta?. An immense amount of variation may occur, moreover, in the notopodial seta? in
any one parapodium.

With regard to the elytra, I have spoken of them as structures which do not belong
to the parapodium. Their attachment is always to the dorsal surface of the somite
proper ; they occur side by side with notopodial cirri in Sigalion and its allies ; and
I have observed in P. (Harmothoe) areolata (I have not examined many other species
from this point of view) that distinct rudiments of notopodial cirri — the whole of the
basal portion, in fact — coexist with an elytron upon the elytron-bearing somite. The
elytra have been spoken of as branchial in function ; my sections (P. areolata) show that
no blood can pass into them. Although they appear to be attached to the somite over a
considerable area, the real attachment is almost linear ; in the central region there is no
attachment, and the epidermis is unbroken. Numerous muscles are attached all round
this ring, and serve to move the elytron as a whole. Nerves pass into the substance of
the elytron (which is quite solid), and, forming a regular network (which I have stained
with gold), end in the numerous small papilla?, which are, no doubt, tactile and similar
to those Claparede has described upon the cirri.

Characters of Polynoe clava (Montagu). — These may be summarized as follows : —
27 somites ; 12 pairs of elytra, borne upon somites 2, 4, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23.
Median tentacle twice as long as the lateral tentacles, bulbous below the apex, with a
terminal filament. Lateral tentacles with similar structure. Palps tapering to a point,
beset with six longitudinal series of small papilla?. Normal neuropodial seta? clavate, with
talon-like apex, 7 or 8 pairs of pectinated frills, the most anterior provided with 5 long
teeth upon its fine edge. Notopodial cirri about twice the length of the parapodium,
with the same structure as the prostomial tentacles. Neuropodial cirri half the length of
the parapodium, tapering gradually to a point, smooth. Elytron only slightly rough on
the surface, margin entire. Anus on the dorsal surface. Pygidiai cirri resemble noto-
podial cirri in structure.

IN THE ANATOMY OF THE POLYNOINA. 355

DESCRIPTION OF THE PLATES.

pro, prostomium. el, elytron.

per, peristomium. pyg, pygidiuiu.

p, parapodium. buc. cr, buccal cirrus.

nt, notopodium. neph, nephridial papilla.

nr, neuropodiura. oc, eye-spots.

cr, cirrus.

Plate XXXIV.

Fig. 1. Polynoe clava, ventral view.

Fig. 2. Polynoe clava, dorsal view, elytra removed : an, anus.

Fig. 3. Polynoe clava, dorsal view, body distended, elytra not meeting in the medio-dorsal region :

an, anus.
Fig. 4. Polynoe clava, dorsal view, body not distended, elytra overlapping in the medio-dorsal region :

an, anus.
Fig. 5. Polynoe clava, peristomial segment: nt. cr, notopodial cirrus; nr. cr, neuropodial cirrus ; not. set,

group of notopodial bristles ; nt. ac, notopodial acicula.
Fig. 6. Polynoe clava, pygidial segment : pyg. cr, pygydial (neuropodial) cirrus ; nt. ac, pygidial

(notopodial) acicula.
Fig. 7. Polynoe squamata, anterior elytron. Owing to an error on the part of the lithographer this figure

has been turned round; the lettering ant., post., ext., and int. indicate its proper position.

Plate XXXV.

Figs. 8 & 9. Polynoe clava, anterior and posterior elytra respectively : at. position of attachment.

Fig. 10. Polynoe clava, portion of elytron.

Fig. 11. Polynoe clava, distal portion of notopodial bristle.

Fig. 12. Polynoe squamata, distal portion of notopodial bristle.

Figs. 12 a, 12 b, 12 c. Variations of same, all occurring in one parapodium.

Fig. 13. Polynoe clava, distal portion of neuropodial bristle.

Fig. 14. Polynoe squamata, distal portion of neuropodial bristle.

Plate XXXVI.

Fig. 15. Polynoe squamata, prostomium, showing eyes and bases of lateral and median prostomial tentacles.

Fig. 16. Polynoe clava, prostomium, showing the same.

Fig. 17. Polynoe clava, ventral view of two parapodia : neph, nephridial papillae,

N.B. Figs. 1S-21 are drawn to the same scale.

Fig. 18. Polynoe clava, nephridial papilla, showing grooves on surface.
Fig. 19. Polynoe squamata, the same.
Fig. 20. Polynoe spinifera, the same.
Fig. 21. Polynoe areblata, the same.

SECOND SERIES. — ZOOLOGY, VOL. II. 50

35(5 ON CERTAIN POINTS IN THE ANATOMY OF THE POLYNOINA.

Fig. 22. Diagram of the nephridium of the Polynoina : ci, ciliated internal opening ; /, lumen of tuhe ;
v, vesicle ; nepk, nephridial papilla ; cut, cuticle.

Fig. 23. Polynoe spinifera, transverse section of nephridial papilla: cut, cuticle; ex. ep. epidermis;
ct, connective-tissue cells ; int. ep, epithelium of nephridial tube ; /. lumen of excretory duct.
x 550.

Fig. 24. Polynoe spinifera, transverse section of nephridial tuhe near the internal opening : int. ep, ciliated
epithelium of wall of tube ; c, granular contents, x 550.

Fig. 25. Polynoe spinifera, transverse section of excretory duct of nephridium in the region of the vesicle,
showing collapse of the walls owiug to the pressure of the ova in the coclom.

Fig. 26. Polynoe spinifera, longitudinal section of a neuropodial cirrus, showing : — ct, branched connective-
tissue corpuscles ; c, large counective-tissue cells ; n, nerve, x 000.

[ 357 j

XI. On Simondsia paradoxa and on its Probable Affinity with Sphseralaria boinbi. By
T. Spencer Cobbold, M.D., F.R.S., F.L.S., Correspondent of the Academy of
Sciences of Pit iladelplt ia.

(Plate XXXVII.)

Read March 15th, 1SS3.

JLlIIRTY years ago (that is, in March 1852) several examples of a remarkable parasite
were discovered by Professor James B. Simonds in the coats of the stomach of a German
hog. The "host" died at the Zoological Society's Menagerie, Regent's Park.

The only notice, hitherto recorded, of the worm occupies a few lines of my intro-
ductory treatise published in the autumn of 1801 ('Entozoa,' p. 79). The notice is
fragmentary and unaccompanied by any technical description of the worm, which,
however, I provisionally named Simondsia paradoxa in honour of the discoverer. It was
only at the last moment, when the said treatise was going to press, that the late Principal
of the Royal Veterinary College called my attention to his specimens and drawings ; hence
the paucity of details, and other deficiencies in the record there given. What, at the time,
was regarded as the head of the worm turns out to be its tail ; and the inference which
Professor Simonds had drawn respecting the strongyloid character of the parasite proves
to be incorrect. In justice to the discoverer, it must nevertheless be said that similar
misinterpretations have repeatedly occurred, as, for example, in the case of Rcederer's
Trichuris, now universally known as Trichocepladits.

Shortly after the original notice appeared, the drawings of Simondsia were lost. The
illustrations had been executed by a skilled artist, the late Mr. Woodman, under Mr.
Simonds's superintendence. Only quite recently, on the occasion of Mr. Simonds
retiring from London, the illustrations were found ; and thus at length both the drawings
and specimens have become available for establishing the genus on a securer basis. As the
specimens are unique and also few in number, I have been reluctant to mutilate them ;
but, fortunately, their transparency is such that nearly all the important parts can
be seen without dissection.

Although in 1864 I had only opportunity to make a brief outward examination,
enough was seen to justify the employment of a new generic name, whilst the specific
title bore reference to the presence of a large spherical and somewhat flattened organ
whose significance was altogether paradoxical. The singular development in question
many times exceeds in size all the rest of the parasite, its diameter being fifteen times
that of the body of the worm at its centre (PI. XXXVII. figs. 1 & 2). In the original
notice this organ was described as consisting of " special folds of integument formed
for the lodgement of unusually developed uterine organs ;" but there is here a false
interpretation. At the time this conclusion was formed I had not examined the organ
microscopically, neither had helminthologists been made acquainted with Schneider's

SECOND SERIES. — ZOOLOGY, VOL. II. 51

358 DR. T. S. COBBOLD OX SMOXDSIA PAEADOXA.

interpretation of the somewhat analogous appearances exhibited by Spheerularia bombi.
I am not sure, indeed, that, even now, Schneider's view meets with universal acceptance.
In Sir John Lubbock's well-known first memoir on Spheerularia (which not only added
largely to our knowledge of the distribution of that parasite amongst bees, but also
extended the facts of its minute anatomy), it was reckoned that the male Sph&rularia
was " twentv-eiffht thousand times smaller than the female." Now, what our President
regarded as the female worm, Schneider avers to be the uterus only, and that wbich our
President not unnaturally regarded as the male parasite, Schneider has declared to be
the female worm deprived of her uterus or, rather, having her uterus, several thousand
times bigger than her body, in a state of prolapsus*. To this point I shall advert again
presently. Meanwhile, with a provisional acceptance of Schneider's view, I think that,
while on the one hand we must regard the genus Simondsia as altogether unique, we
must, on the other hand, recognize its approach towards Spharularia in respect of the
enormously developed reproductive organs of the female, which, in both genera, lie, as <
it were, outside the body proper.

Until the appearance of Sir John Lubbock's first memoir, in the 'Natural History
Review,' the so-called male Spheerularia had never been indicated. If Schneider's views
be correct, it obviously results that the male worm is still unknown to science, and that
its discovery, as announced in the pages of the 'lleview' for 18G7, really refers to part
of the body of the female Spheerularia : in fact, her uterus was the only part hitherto
seen by Dufour and other observers. When, in 1^(57, Professor Huxley lectured at the
Royal College of Surgeons, I understood him to accept Schneider's view without hesi-
tation; nevertheless, in his more recently published 'Manual,' where he deals with the
<;enus Spharularia, no reference is made to the Berlin helminthologist's opinion. I was
especially interested in this question, inasmuch as I had, during the previous year,
prepared and mounted several examples of Spheerularia bombi lor the College Museum f.
From the abdomen of a bee, which was one of the old Hunterian specimens, I removed
three Spharulariee, and, in accordance with Sir John Lubbock's views, described them as
" three females, two of them having extremely minute males sexually attached." The
catalogue was prepared in 18G5, nearly a year prior to the publication of Schneider's
well-known monograph.

During the month of May, ISO.", the late Dr. Ormerod, of Brighton, sent me
some parasites from Vespa rufa and Vespa vulgaris. These I at once recognized as
examples of Spheerularia ; and I noted the tact that all the insect hosts were females.
This peculiarity of habit agrees with what Sir John Lubbock has observed in the various
species of infested Bombi and in Apathus. Now, whatever be the significance of this
singular habit, it is one which is by no means confined to the aberrant genus of jiarasites
in question. The same peculiarity affects Filaria terebra which infests the abdomen of
the female black-tailed deer [Cervus columbianus).

'fhe circumstance that the male Simondsia paradoxa (PL NXXVII. fig. 3) exhibits

* Monographie der Nematoden, 1866, S. 322.

t Catalogue of Eutozoa in the- Museum oi the Royal College of Surgeons (London, 1SG0), Nos. 37, 38, p. fi.

DR. T. S. COBBOLD OX SDIOXPSIA PAEAPOXA.

no departure from the ordinary nematode type of structure prepares us. in some degree,
for a similar outcome in respect of the hitherto undiscovered male SpJuerularia. Be
that as it may, the mere fact that Sir John Luhbock did not notice any spicule in the
jody of the alleged male Spharularia is pretty strong proof that the minute creature
.yas not of that sex. Professor Schneider at once saw the significance of this negative
character *.

The chief point of interest in Simoadsia lies in a correct interpretation of the morpho-
ogy of the organ which I call the " rosette." I do not pretend to have settled the
question of its homology ; but, guided by the facts observed, I think that the rosette is
a prolapsed uterus. It is thus the equivalent of the so-called " generative animal" of
Spharularia, whilst the body proper represents the " nutritive animal," as interpreted
jy Schneider. It is furnished internally with two egg-containing horns. The rosette
-hows evidence of spiral twisting ; and this may sufficiently account not only for its
flattened spherical figure, but also for its unlikeness, in this respect, to the equi-
valent organ in Splucrularic. I presume that the branched external pr( 5S - are homo-
logous with the spherules of Sphantlaria, whilst the ultimate caecal capsules have, as
far as I know, nothing comparable to them in nature. Their outer surfaces are finely
tuberculated. Internally, the cseca and branches of the uterine wall are lined with
granular debris, which I believe to be the remains of cells, comparable to the polygonal
cells figured both by Schneider and Sir John Lubbock in Splnerularia, such cells being
held by the former observer to be characteristic of uterine structure in this class of
animals.

No diagnosis, either of the genus or species, having been given. I submit the
following : —

SnioNDsiA. Cobb.

A genus of endoparasitic nematodes, iu which the female is furnished with au external and much
enlarged uterus, whose walls expand into branches terminating in cava. Females encysted. Males
free.

SlAIOXPSIA PABADOXA, Cobb.

Head bluntly pointed; neck with narrow lateral wings ; mouth simple, with two pro-
minent side papillae; body of uniform thickness generally, but externally supporting, in
the female, a large rosette-shaped organ formed by the uterus ; tail of the male spirally
twisted, suddenly narrowed to a blunt point ; spicules two. long and slender ; tail of the
female twice as thick as the body, conical, bluntly pointed, with three broad-based spines
immediately above the anus.

Length of the male half an inch. Length of female six tenths ot an inch.

Sab. Stomach of the domestic hog (Su$ scroti', var. domestica) ; males free ; females
encysted within the walls, their heads projecting into the cavity of the organ through a
narrow passage of intercommunication.

* Since this paper was read. Trot. Lnton S hneider has announced the actual discovery of the male SpJuerularia
and has cleared up the mystery attaching to its development. See ' Zoologische Beitrage,' Bd. i. S. 1. Taf. 1.
Breslau. 1- .

360 DR. T. S. COBBOLD ON SIMONDSIA PA11ADOXA.

The host was a " boar" suffering from an affection of the skin, a circumstance which led
to its being killed and employed for dissection-purposes. Though well fed, it was a poor,
lean animal, and, during life, afforded neither sign nor symptom of gastric disorder.
Before Professor Simonds detected the parasites, some portions of the stomach had
already been removed ; and thus it is concluded that many specimens of the parasite
escaped observation.

The cavities, or cysts, formed in the walls of the stomach by the female parasite are
remarkable structures. Each gastric cyst contained a single female worm. The cysts
projected both externally, at the peritoneal surface, and internally, causing a bulging of
the mucous membrane.

Professor J. B. Simonds states that there was no evidence of the existence either of
congestion or of inflammation in the surrounding tissues. When a cyst was opened, some
clear fluid escaped ; nevertheless a small aperture at the mucous surface permitted the
head and neck of the female parasite to project into the cavity of the stomach. The
cavity of each gastric cyst not merely enclosed the body of the worm, but formed, as if
were, a perfect cast of the rosette-like organ. On transverse section, each gastric cysf
displays six or seven distinct circular rows of depressions internally (PI. XXXVII. fig. 4).
These depressions are regularly arranged, and are themselves made up of smaller cavities,
generally from four to six in each of the larger depressions. The ultimate ca:cal ends
of the rosette are received into still smaller cup-shaped cavities proceeding from the
secondary depressions. The number of these caeca varies considerably, some of the main
branches of the, rosette supporting upwards of forty of them.

The formation of these peculiar gastric cysts can only have been brought about by the
gradual development of the enclosed parasite, the gentle and equable pressure of the
slowly expanding body of the worm causing so perfect an adaptation of the environment
that every part of the thin-walled rosette became protected from injury. It is hardly
likely that the ultimate purpose of this cyst is concerned with the mere steadying, 01
fixing of the body of the worm during sexual congress.

The symmetry of form and sculpturing of the cysts thus produced by Simondsia is«
almost as perfect as the somewhat similar appearances seen in fossil casts ; and, although
their mode of formation must have been essentially different, it is impossible that the
analogy should escape notice. Prom the helminthic standpoint, these gastric cysts musl
be pronounced unique ; nevertheless, in respect of their mode of formation, something
of the same kind may be observed in the cast-like depressions formed by the anchored
heads of various species of Taenia and Echinorhynchi. Pentastomes and Gastrophilm
larvae also produce depressions of a simple character.

As regards the anatomy of Simondsia, I have only further to state that the integumem
is everywhere striated, except at the surface of the rosette. Only on the tail of th(
female worm were there any spines observed. They occur on the ventral surface, at i
distance of about -^ of an inch from the tip. The spines, three in number, are placet
in a transverse row. They are nearly as broad at the base as they are long, separatelj
measuring about 3^ of an inch from base to apex. The body of the male worm afford;
an average diameter of -£$ of an inch, the female being one sixth broader ; the tail of th<

DK. T. S. COBBOLD ON SIMONDSIA PARADOX A. 361

female, however, at a line corresponding with the spines, gives a breadth of fully -£% of
an inch. The intestinal tract is simple in both sexes, and terminates close to the end of
the tail ; but in the female the gut is much widened within the caudal enlargement of
the body. In either sex, the oesophagus is of great length, measuring fully ^ of an inch
from the mouth to commencement of the chylous intestine. The lateral oral papillae
give a length of about 3-^ of an inch from base to apex. As regards the reproductive
organs of the male worm, we find two long, elastic, and very delicate spicules, each
measuring about the -£g of an inch in length, and only little more than the xcfoo of
an inch in breadth. In the male, both the intestine and the sexual organs terminate
in the usual manner. In the female the whole mass of the ovarian filaments, together
with the tuba? and uterine branches, are lodged within the rosette ; but where the vulva is
situated could not be ascertained with certainty. The junction between the tuba? and
horns of the uterus is particularly well marked (PL XXXVII. fig. 2, a) ; but the mode
of union of the uterine horns with the rosette was not observed. Although no vulva was
actually seen, it may be surmised to terminate somewhere at the base of the rosette in the
ventral line.

The eggs of Simondsia are small in size and extremely numerous (PI. XXXVII. fig. 5).
Always more or less elongated, their figure varies considerably, some being oval, others
elliptical, with a tendency to become constricted at one or both sides. Some of them
finally assume either a reniform or even a more or less hourglass-shaped figure
(PL XXXVII. fig. 6). Their average length is from g^y to ^}q of an inch, their breadth
at the centre in constricted specimens being sometimes less than the ^"oVo °f an inch.
Their walls display a double contour ; and the more advanced of them contain imperfectly
developed embryos.

DESCPJPTION OF PLATE XXXVII.

Fig. 1. Female Simondsia paradoxa removed from its cyst : a, head; b, b, lateral aire; c, lower end of
oesophagus; d d, chylous intestine ; e, e, e, cseca of the rosette; /, tail, x 12 diameters.

Fig. 2. Detached rosette, with horns of the uterus partly drawn out from the interior : a, a, junction of
the tubse with the upper ends of the horns ; b, b, b, uterine branches, x 12 diam.

Fig. 3. Small male Simondsia paradoxa : a, head ; b, tail. xl:' diam.

Fig. 4. Section of the wall of the stomach of Sus scrofa, showing the interior of a parasite-bearing cyst
exposed by transverse section : a, lower half of the cyst ; b, b, b, secondary and ultimate depres-
sions, x 6 diam.

Fig. 5. Fragment of one of the uterine horns, showing numerous ova in the interior. x 150 diam.

Fig. 6. Group of eggs, removed from the above, x 250 diam.

Fig. 7. Outline of the tail of the male worm : a, anus ; b, spicules, x 32 diam. Outlined with the aid
of a camera. Original.

SECOND SERIES. — ZOOLOGY/, VOE. II. 52

[ 3C3 ]

XII. On the Testis o/"Limulus. By "W. B. S. Benham, Esq. (Communicated by
Prof. E. Rat Lanejestek, II A., F.L.S.)

(Plate XXXVIII.)

Read June 21st, 1SS3.

WHILE working in the laboratory at University College, last October, as one of Prof.
E. Ray Lankester's assistants, he suggested that I should examine the testis of the King-
Crab. By his kindness (here I will take the opportunity of thanking him for his kind
help and suggestions) I was fortunately able to work on the fresh animal. Owing
to the difficulty, and the time taken, in removing the carapace, not much could be
done in a day ; and though it kept fresh till the next day, after that it had to be placed
in spirit ; in this state the finer networks are easily broken, and the finer blood-vessels
may be mistaken for the ducts, though the mistake is soon discovered, owing to the
absence of networks in the former. The testis is packed by the ramifications of the
gastric gland (the so-called liver), and lies almost superficially.

The only description of the generative organs of the King-Crab appears to be that by
Prof. Owen, which was published in the Linnean Transactions for 1873. He there
describes and figures a portion of an ovary ; his description opens thus : — " The ovarium
is a system of ramified tubes and cavities, occupying chiefly the dorsal region of the body ;
it extends along the median part of the thoracetron" ( = abdomen), " and expands laterally
in the cephaletron " (cephalothorax). His figure shows rather thick ducts; and the network
represented is not extensive, but, so far as it goes, corresponds roughly with the distri-
bution of the testis. It may be said that the testis forms a reticulated tube all over the
dorsal portion of the thorax and abdomen, mixed up with the liver. Owen divides the
ovary into anterior and posterior median lobes, with anterior, lateral, and postero-lateral
branches and networks proceeding from the former lobe.

We can scarcely, as will be seen, divide the network of the testis into such lobes, though
roughly there are anterior (thoracic), posterior (abdominal), and lateral (thoracic) net-
works. As in the ovary, the main duct of the testis, from the external aperture, divides, on
reaching the surface of the liver, into three main branches, having about the same course
as in the female organ.

Professor Owen's figure represents the branches as symmetrical on each side ; such is
' not the case with the network of the testis, as will be seen. He suggests as a cause for
the posterior median lobe of the ovary dividing into two branches, at the junction of
abdomen and thorax, which run forward to the anterior median lobe, that the pressure
on it, between heart and intestine, which would take place on flexion of the abdomen on
the thorax would be too great if median, and hence there is a branch on each side, out-
side the pericardium. In the case of the testis, the ducts are much smaller than those

SECOND SERIES. — ZOOLOGY, VOL. II. 53

364 MR. \V. B. S. BENIIAM ON THE TESTIS OF LIMULTTS.

represented in his figure of the ovary, so that they occupy but little space between the
pericardium and intestine, and the pressure on it would be very slight. The external
aperture, both of oviduct and sperm-duct, is similarly situated on the posterior face of the
vii. appendage, about one third of the distance from its attachment, near the middle
line on each side (as shown in fig. 2).

The course of the spermatic network. — The sperm-duct rises from the external aperture
on the vii. appendage, upwards and outwards, parallel to the "anterior lamellar"
muscle of this appendage. It reaches the surface of the liver, through which it runs,
just within the muscle attached to the sixth thoracic limb (tig. 1, D1, Dr).

Lateral network. — Here it breaks up into three branches ; one runs backwards and
outwards, behind this large coxo-tergal muscle, giving off branches, which anastomose
with one another, forming a network, which includes in its meshes some of the smaller
muscles of the sixth coxa ; the ducts get finer as they recede from the main duct (they are
drawn too thick in the figure), and they run to the side and to the posterior corner of
the thorax.

Just after passing the large muscle, an anterior duct is given off, passing along the
outer edge of the coxo-tergal muscles, forming a network, which probably runs right away
to the front ; but I was unable to follow it further than is represented in the figure, as
here the ducts get very fine and easily break. The second main branch from the duct
runs inside the coxo-tergal muscles, forwards, giving off networks between the various
muscles, some of the smaller ones being included in the rncshes ; these networks pass
outwards, and probably anastomose with that on the outer side of the coxo-tergals.
However that may be, inside it runs right away in front of the muscles, and on its inner
side gives off branches to a median network.

The median nehcork. — The third principal branch is represented on the right side, and
does not appear on the left, so that there is a certain amount of asymmetry. This runs
inwards at first, behind the branchio-thoracic muscle, then forwards along its inner border,
where it gives off a network across the middle line, lying between the pericardium above
and the alimentary tract below. This network {A, fig. 1) was removed, and cleaned, as far
as it could be, from liver and connective tissue, and mounted ; it is represented in fig. 3,
where A A represent the lateral boundaries ; on the ducts are seen the sperm-sacs,
which are usually in groups, one of which opens into the duct, and the others into one
another.

In front of this network is another continuous with it, but separated from it in the
figure, owing to the rupture of some of the finer ducts. The anterior lateral boundary of
the left side is formed by a duct running from the front of the left branchio-thoracic
obliquely forwards to the right side. The network alongside the second coxo-tergal muscle,
the most anterior of the large muscles, sends small branches forwards, but not very far,
and backwards a larger duct, whence branches and networks pass between the muscles,
probably to join a similar lateral network on the right, as was described for the left side.
The left border of the median network is formed by two very fine ducts parallel with
one another, on the inner side of the branchio-thoracic muscle. These two ducts anas-
tomose here and there, and join a network in the abdomen, across the middle line, beneath

ME. W. B. S. BENHAM ON THE TESTIS OF LIMULUS. 365

the pericardium ; the ducts theu become excessively fine, and the network, at any rate
the sides of it, ends at about the xi. appendage. The network was here incomplete, as I
had removed the portion A previously to noticing these fine posterior ducts.

The lateral network of the right side is doubtless similar to that on the left ; but I did
not follow it throughout its extent.

Thus there is a continuous network of ducts more or less all over the body, lying, in
the median line, below the pericardium and above the intestine; this median network
runs throughout the abdomen and thorax, though the ducts are larger more anteriorly.
This corresponds to Prof. Owen's "anterior and posterior median lohe" of the ovary;
the lateral lobes or branches of this are represented by the lateral networks outside the
coxo-tergal muscles of the thorax.

That the testicular network does thus extend everywhere is shown by the fact that,
when the carapace is cut off, a white viscous fluid flows out, which consists of blood mixed
with spermatozoa ; this soon coagulates.

The spermatozoa consist of a small rounded head about two iuch in diameter, with a
flattened wide " neck " and a long mobile tail (fig. 1). The mobility of the spermatozoa
of Limulus was discovered and recorded by Prof. Lankester in 1881 (Quart. Journ. Micr.
Science, vol. xxi.).

The sperm-sacs are seated in groups upon the sides of the ducts ; two or three, some-
times more, are in connexion with one another, one of which opens into the duct (figs. 3
& 5). These sacs are subglobular in shape, and are surrounded by a delicate membrane.
These are filled with groups of spermatozoa without tails, their further development
apparently taking place in the ducts as they approach the aperture. Sometimes sperm-
sacs are seen at some distance from a duct, and I could, in many cases, trace no ductule
from them (as at B, fig. 5). Perhaps it is here that the sperm-cells are produced, and
by continual enlargement of the sperm-sacs, by subdivision of the sperm-cells, and
stretching of membrane round them they reach the ducts.

All these structures, as well as the caeca of the great gastric gland, are packed by a
fibrous connective tissue, which is very loose (fig. 5, C).

The wall of the duct (fig. 6 ) is made up of interlaced tendinous-looking fibres ; by
tendinous I mean that they are more or less highly refractive and parallel to one another.
These are arranged longitudinally, circularly, and obliquely, while the inner surface is
lined by columnar cells. The duct is often filled with the unripe spermatozoa, which,
here and there, have aggregated into rounded masses near the walls (fig. 5, S).

Significance of the above observations. — The importance of the observations thus
recorded depends upon the fact that in no crustacean do the ducts of the generative
glands form a network, whereas in the Scorpion (and other Arachnids), as in Limulus,
they do. Hitherto the fact that the ovarian ducts of Limulus form a reticulum was
the only information which we possessed as to the form of the generative organs in
Limulus. The observations recorded in the present paper show that the duct of the
male organ in Limulus also has the form of a reticulum. Thus a new and important
confirmation is afforded to the proposition maintained by Prof. Lankester in his essay

366 ME. W. B. S. BENHAM ON THE TESTIS OF LIMULUS.

" Limulus an Arachnid," viz., that the organization of Limiilus is closely similar to tka
of the Scorpions, and has no special agreement with that of any Crustacea.

A second point of considerable interest is the apparent isolation of many of tk
spermatic sacs, and the probability that they arc not diverticula of the spermatic duct
but, secondarily, in the course of their ripening acquire a connexion with it, the twe
structures having developed independently. This is in accordance with what is knowr
as to the essential nature of the genital ducts of Arthropoda, although it cannot be saic
that we have, at present, a sufficient series of observations upon their development tc
warrant a generalization.

DESCRIPTION OF PLATE XXXVIII.

Fig. 1. The carapace lias been removed from the cephalothorax (Ceph), and the tergum from the
abdomen (;lbd), to show the course of spermatic network amongst the liver (//'), which is not
indicated: A, median network ; L, lateral network ; Cxl-Cx°, coxo-tcrgal muscles, the rounded
knobs (k) to which these are inserted are parts of the cntocoxites of the appendages ; Br.lh,
branchio-thoracic muscle ; I)1, J)r , the hit and right main sperm-ducts, dipping down to the
external aperture in appendage vii.

Fig. 2. Genital operculum, half nat. size. Posterior face, to show the papillae in which the external
apertures of the genital duct arc situated ; gen, d, sperm-duct dividing into three main branches ;
g.p, its external aperture; Stg, stigma; t.s, tendinous stigmata; ant., anterior lamellar
muscle; ext., external branchial muscle.

Fig. 3. Portion of the median network A (fig. 1), magnified about £H times; AA, lateral boundaries of
this portion ; d, duct ; sp, sperm-sacs.

Fig. 4. Mature spermatozoa.

Fig. 5. Portion of a section showing sperm-sacs, x 22 : d, sperm-duct ; W, wall of duct ; L, epithelium
of duct; -S, aggregations of spermatozoa ; c, surrounding connective tissue; M, membrane of
sperm-sac ; B, sperm-sac with no apparent connexion with duct.

Fig. 6. Portion of wall of sperm-duct, seen from without ; enlarged about 39 times.

Fig. 7. Transverse section of duct, x39: d, lumen of duct; L, epithelium; n, nuclei; W, wall; S,
spermatozoa within the duct.

[ 367 ]

XIII. The Metamorphosis 0/ Filaria sanguinis hominis in the Mosquito. By Patrick
Manson, M.D., Song Kong. Communicated by Dr. Cobbold, F.B.S., F.B.S.

(Plate XXXIX.)

Eead March 6th, 1884.

>IX years ago I described the metamorphosis undergone by the embryo Filaria san-
guinis hominis in the body of the mosquito *. I hoped that (considering the practical
importance of a correct knowledge of the life-history of this parasite) the statements I
:hen made would, long ere this time, have been thoroughly confuted or confirmed. If
we judge from the wideness of its distribution and the grave character of the diseases it
rives rise to, the Filaria sanguinis hominis is infinitely the most important of the animal
parasites attacking man, much more important than Trichina or Echinococcus. Bio-
logically the story of its life- history is interesting, and even more wonderful than is that
of either of these parasites. Nevertheless, although from both a practical and a scientific
point of view it is important to have this assertion of mine about the role of the mosquito
cleared up, it still remains but half confirmed, half believed.

With the exception of Lewis in India f, Myers in Pormosa %, and Sonsino in Egypt §,
I do not know that any one has worked seriously at the subject. And although both
Lewis and Sonsino have confirmed my statements as to the entrance of the Filaria into
the mosquito, and followed up part of the metamorphosis, neither of them has advanced
his observations so far as to be able to confirm my statements as to the later stages of
this, or positively to prove that the mosquito is, or is not, the intermediary host.

Some eminent helminthologists in England accept my statements and endorse the
inferences I have drawn — Cobbold for example. But in other quarters, so far from
securing acceptance of my theory, the work of Lewis, on account of the hesitation and
scientific caution with which he expresses himself, has had the effect of inducing a
certain amount of scepticism. Leuckart is sceptical; and of course the scepticism of
so eminent an authority is of great weight in influencing opinion, especially in Germany.
Some of our own zoologists, also, I understand, share the views of Leuckart.

Recently I received a copy of a lecture by Mr. B. Scheube of Leipzig entitled ' Die
Pilaria Ivrankheit.' In this very excellent lecture occurs the following passage : — " Der
englische Parasitolog Cobbold hat Manson's Ansicht von der Entwickelung der Filaria
sanguinis hominis adoptirtund auf die Analogie mit der Filaria medinensis hingewiesen,
welche ihre Metamorphose im Siisswassercyclopen durchmacht. Von Leuckart ||
dagegen werden gegen die Manson'schen Beobachtungen aus mehreren Griinden starke

* Proc. Linn. Soc. March 7th, 1S78. China Customs Medical Eeports, Sept. 1877.
t 14th Ann. Eep. of the Sanitary Commissioner with the Government in India.
+ China Customs Medical Reports, March 1881. 21st issue.
§ Medical Times and Gazette, May 13th, 1SS3, p. 404 ; Sept. 22nd, 1883, p. 340.
|| ' Die Parasiten des Menschcn,' I. 2. Aufl. p. 85. Leuckart's Jahresbericht fiir 1875-70, p. 163.
SECOND SERIES. — ZOOLOGY, VOL. II. 54

368 DE. P. MANSON OX THE METAMORPHOSIS OF

Zweifel erlioben, welohe audi die spiiteren Mittheilungen von Lewis nicht beseitigen
konnten. Letzterer constatirte zwar ebenfalls, class die Filaria sanguinis hominis in den
Magen der Muskitos iibergeht und bier gewisse Veranderungen erleidet ; er fand anch
bei denselben jugendlicbe Nematoden, die moglicher Weise einen weiteren Entwickelungs-
zustand der aufgenommenen darstellen, aber es ist durchaus nicbt sicber, dass dem in
Wirklicbkeit so sei. Von einer anderen Seite konnten die Manson'scben Beobacbtungen
nicht bestiitigt werden. Myers wollte dessen Versucbe anf Formosa, wo die Filaria-
Krankbeit selbst nicbt autochthon vorkorarat, nacbmachen, kam aber zu dem Resultate.

dass die Filaria-'Embryoneia. von den Muskitos vollstandig verdant wurden

Nach alledem miissen wir die Frage iiber die Entwickelung der Filarie noch als ein(
schwebende betracbten." These words, I presume, represent the attitude of the
scientific and professional mind, on the continent at all events, on the subject of Filaria-
metamorphosis. My assertions may be true, but in all particulars they have not beer
confirmed, and the evidence I formerly adduced cannot be considered sufficient.

Impressed by the practical importance and scientific interest of this matter, and being
naturally anxious that what I had stated and knew to be true should be accepted, and '
failing decided confirmation from other observers, I thought I could do some service tc
my profession and this branch of helminthology if I again went over the ground I bad
twice before trodden. With these and other criticisms before me I have done so, anc
now endeavour by a short statement of what I found, once for all to settle the matter.
This statement I supplemented with a series of illustrative drawings carefully made from
fresh preparations and drawn to scale. And, in order that I may quote the testimony
of others, I have shown the whole series of my experiments and preparations to Drs.
Macleish and McDougall, medical men residing in Amoy, and perfectly competent judges
on such points. I have their permission to say that they believe my descriptions and
drawings to be in the main correct. Further, I have forwarded to friends in England
and elsewhere microscopic preparations which, should the frail structures they contain
retain the appearances they had when they left my hands, cannot fail to satisfy the most
cautious and scientific mind.

There are three ways by which we may settle the relationship of the mosquito to the
Filaria : —

1st. By tracing the Filaria into this insect, witnessing its metamorphosis therein, and
finally its escape at an advanced stage of development.

2nd. By showing that it is only in those mosquitoes that have fed on filariated blood
that the metamorphosis of a nematode occurs — that no such phenomenon can be witnessed'
in mosquitoes fed on non-filariated blood.

3rd. By filariating a man, or a lower animal, by means of Filaria? metamorphosed in
passing through the mosquito.

The first two of these methods I have employed ; the sequel contains a description of
my results. The third test of my theory remains to be applied. For myself I have not
sufficient hardihood or scientific fervour to attempt it in my own person, but I am con-
vinced that a properly planned and conducted experiment would lead to positive results.
Both Lewis and Sonsino, besides other observers, confirm me in the statement that the

FILAEIA SANGUINIS HOMINIS IN THE MOSQUITO. 3G9

mosquito, when feeding on the blood of a filarious subject, imbibes the Filaria. Lewis
confirms me in the statement that nematodes at a stage of development slightly more
advanced than those found in the newly ingested blood are also to be found in the
mosquito. But Lewis says he could not satisfy himself that these two forms belong to
one and the same species of nematode. Neither he nor Sonsino has found advanced
forms, so advanced that from their appearance they would be justified in concluding they
were fit for independent life. Possibly, suggests Lewis, the more advanced forms are
embryos of some other species of nematode, and not the Filaria sanguinis hominis at all.
In my observations I paid particular attention to this point. If I succeed in connecting
the more advanced forms of nematode embi-yo, as seen by Lewis, with the unquestioned
embryo Filaria sanguinis hominis found in newly ingested blood, and with the later forms
which I had already described as being equipped for independent life, then there is no
longer any room for doubt about the role of the mosquito, and I can claim that Lewis's
observations confirm my own.

I would remind the reader that the parasite called Filaria sanguinis hominis is a minute
nematode embryo which in certain warm countries is found in the blood of Man. It is
the offspring of a mature Filaria living in the lymphatic vessels. The young find their
way into the blood-circulation along with the lymph.

No form intermediate between the mature Filaria and its embryo has hitherto been
found in Man. There is no evidence of growth about the embryo, neither has any
provision in its structure by which one might suppose it could pass from one man to
another been discovered. It seems most probable, therefore, that like many other
parasites it requires the services of an intermediary host — 1st, to remove it from the body
of the host ; 2nd, to nurse it till it becomes equipped for independent life ; 3rd, to place
it in a position for obtaining access to its final host.

In our search for this intermediary host we must be sure the animal we encounter can
fulfil these conditions. In addition to this there are two other circumstances in connex-
ion with the Filaria which must also be recognized in fitting it with an intermediary
host. Such an animal must have a geographical distribution corresponding to that of
the Filaria ; and it must also be nocturnal in its habits. The reason for the first of these
conditions is self-evident ; the latter is indicated by the curious fact in the history of the
Filaria, that the embryo appears in the blood only at night. This peculiarity in the
habits of the Filaria has now been so frequently demonstrated and confirmed that I need
concern myself here only with the mention of it.

There are many reasons why one should regard the mosquito as the animal most likely
to fulfil all the conditions required. These reasons I will not enlarge on. I will confine
myself to the description of what actually occurs in the relationship of the mosquito to
the Filaria, feeling satisfied that in this is the best proof that the mosquito is the true
intermediary host of the Filaria sanguinis hominis.

In experimenting with the mosquito there are two points I was careful in attending
to ; I employed the proper species of insect, and I took care that its only food was the
blood of a filariated man. Those who would repeat my experiments must bear these two
points in mind. The plan of procedure I found most convenient was as follows : — I

54*

370 DE. P. MANSON ON THE METAMOEPHOSIS OP

engaged a man in whose blood the Filaria abounded. A large square wooden frame
(10 ft. x 10 ft. x Gh ft.) covered with mosquito-netting and provided with a door was
erected in a room where mosquitoes abounded. Under this mosquito-house the man
slept, the door of the house being closed some hours after he had gone to bed. Next
morning the mosquitoes that bad entered by the door, and preyed on the man's blood,
were found clinging to the inside of the netting, their abdomens distended with blood.
These my servant captured under a wine-glass, and after paralyzing them with a whiff of
tobacco smoke transferred them each to a dry phial covered with gauze. When the
insects had recovered from the effect of the tobacco and ascended the side of the bottle
for a little distance, a small quantity of water was introduced through a pipette. Latterly
1 employed large-mouthed stoppered bottles, holding about two ounces, instead of the
phials. Tobacco could in this way be dispensed with, and a considerable mortality
avoided. The bottles were dated and put under a dark shade. When the weather
became cool I got better results by placing the bottles in an incubator, where the insects
were kept in a damp atmosphere at a temperature ranging from 80° to 85° Eahr. In
this way I endeavoured to imitate nature. Every morning during many weeks a fresh
crop of Filaria-chargeA mosquitoes was gathered. Eor the purposes of this investigation
quite a thousand insects had to be provided. From time to time, as required, they were
removed from the bottles after being killed by chloroform. The legs and wings and head
were torn off. The thorax was then separated from the abdomen, each being placed on
a separate slide. In the case of the abdomen the contents were expressed by rolliug a
penholder from the free to the severed thoracic end ; a cover-glass was then placed on
the expressed blood. If necessary, a little water or sulphate of soda solution (sp. gr.
1050) was used to soften the blood and allow of the easy removal of the two large ovisacs,
which, when crushed, obscure the slide very much. The thorax is best treated by being
broken up, and teased out with a couple of sharp needles in a droplet of water, before
the cover-glass is applied. Thus prepared, an inch objective readily detects the parasites
among the tissues of the insect. To preserve the Filarice thus displayed the cover-glass
must be carefully lifted up, and what remains on the slide allowed to dry, the large pieces
of mosquito debris being first picked off with a fine forceps. After drying for a day or,
better, two, it may be stained with an aniline dye (gentian violet answers very well), and
after washing and redrying mounted in Canada balsam. In the case of Filaria in the
blood from the abdomen of the mosquito it is only necessary to dry the preparation
before mounting in balsam. The red colour of the blood gives sufficient contrast to
display, at all events, the outline of the parasites. Whenever practicable, fresh and
unstained preparations should be examined ; for while stained specimens show the outline
well enough, they are useless in studying the details of internal structure.

My mosquito-house and filarious patient were visited by three species of mosquito,
perhaps four. All of these were capable of imbibing the Filaria, but only one of them
is, I believe, capable of conducting the metamorphosis to a successful conclusion. Both
of the impotent species are of the kind known as tiger mosquitoes, i. e. their bodies and
legs are banded, or flecked with white. One is quite a large insect, half an inch in
length, with numerous sooty, easily detached scales ; the eggs of this species when

FILAEIA SANGUINIS HOMINIS IN THE MOSQUITO. 371

deposited on the surface of water, do not keep together in a hoat-shaped cluster, as do
those of the true Filaria mosquito, hut float about separately, tending to adhere to the
sides of the bottle. There is no danger of confounding this species with the true Filaria
mosquito. The other species of tiger mosquito is smaller, and in this way is liable to
be mistaken for the Filaria mosquito ; but a close inspection shows it to be smaller in
size and darker in colour than the true species ; and, in addition to these points of
distinction, the little flecks of white and white bands on legs and body serve, with a
little care, to prevent mistake. I have often found the Filaria in the blood in the
abdomen of both of these mosquitoes, but never in such numbers as one finds them in
the true Filaria mosquito. In the thorax I have found them slightly advanced in
development in the larger tiger mosquito ; but although I have examined a considerable
number of this species on the third day after they had fed, I have not found the Filaria
at this later period, when in the true Filaria mosquito they abound, and cannot readily be
overlooked. I would not be positive on this point, as my observations on the tiger
mosquito are not sufficiently numerous ; but my impression is that both species are
incapable of acting as efficient intermediary hosts.

The true, or what I believe to be the true, Filaria mosquito is a brown insect, about -^
of an inch in length, snuff-brown in colour, without any particular markings whatever.
In some specimens, and especially some days after feeding, and when her eggs have been
deposited, there is an appearance of banding about the abdomen ; but there are no
distinct markings, as in the case of the tiger mosquitoes. It appears to thrive best when
the thermometer ranges from about 75-84° Fahr. In the very hot weather it dis-
appears, or is languid, and the rate of maturation of the ova (in those specimens one can
procure) is too quick apparently for the metamorphosis of the Filaria. Consequently,
experiments in June, July, and August may not be uniformly successful. In Amoy,
September and October are the best months for observations on this species. The
female alone is the blood-sucker. She feeds but once if not disturbed, and lavs her egg
about the fourth or fifth day afterwards. Some lay on the third clay, and some do not
part with their ova till the seventh. As a rule, they die soon after laying, but a few
survive till the sixth or seventh day, and in these alone is the metamorphosis of the Filaria
completed. Perhaps one mosquito in ten attains this age, at all events in confinement.
Possibly in normal surroundings a larger proportion survive.

On commencing the study of the metamorphosis of the Filaria in the mosquito the
first circumstance that strikes one is the largeness of the number of Filaria? ingested by
the insect. In a corresponding quantity of blood, drawn in the usual way from the
finger, not one fourth part of the number contained in the blood of the insect's abdomen
would be found. It is quite a common thing to find 30 or 40 Filarial in the minute
speck of blood we express from the abdomen. Often this number is exceeded, and more
than once I have found them literally in hundreds. This indicates that the mosquito
in some way or another is adapted for fishing the Filaria out of the blood stream. The
proboscis of the insect, as it lies in the lumen of the vessel from which blood is being
imbibed, must tend to arrest the parasites as they are swept against it by the stream,
just as a stake in a stream of water accumulates straws and sticks ; and the lashing

372 DR. P. MANSON ON THE METAMORPHOSIS OF

movements of the little animals tend to entangle them still further, and bring them
under the influence of the suction-force exercised by the insect.

Arrived in the abdomen of the mosquito the Filaria for a short time retains the appear-
ances and movements that it exhibited while in its human host, and which have
frequently been described (fig. 1). Presently a delicate, closely set, transverse striatiou,
as if from general longitudinal shrinking, gradually becomes very evident. The sheath
also in many cases can be more readily seen, and oral pouting is very distinct. Within
an hour of the time of ingestion, the Filaria usually casts its sheath, which, by careful
searching, may be seen either trailing after it, or lying across it, or somewhere near.
After the casting of the sheath transverse striatiou and oral pouting become still more
distinct, and one can almost fancy the parasite has a true mouth surrounded by lips.
It is impossible to determine this, however.

Until the casting of the sheath, the peculiar lashing movements of the animal continue ;
but when this has been effected a change in the character of these movements is in many
instances observable. Hitherto they have been free, lashing, purposeless apparently;
and although the little animal has kept in perpetual and vigorous motion, it has never
changed, materially, the spot it moved in. There has been no locomotion. But when
the sheatli has been got rid of, the nature of the motion changes to a suake-likc wriggle
of regular undulations, which cause a definite forward movement (fig. 2). In many
specimens this is very striking. If motion is prevented by some insuperable obstacle
in one direction, the Filaria retreats and tries in another. It is evidently endeavouring
to change its locality.

The object of this singular change in the character of the movements exhibited by
the Filaria is explained if we examine the thorax in a batch of insects at short intervals
after they have fed. We tease it up with needles in a little water, or sulphate-of-soda
solution, tearing up the muscles of the wings thoroughly, before applying the cover glass.
Examined, thus prepared, within a short time, say half an hour, of feeding, all the Filarice
are to be found in the abdomen, none in the thorax ; after an hour, two or three may
be found moving among the muscles of the thorax ; after another hour, many more may
be detected here, till at the end of 12 or 18 hours, the thorax is found to be full of
parasites. Therefore, the movement of progression we see in the Filaria that has cast
its sheath in the blood in the abdomen, has for its object the migration of the parasite.

Not all, but a very large proportion of the injected Filaria do thus migrate. My
former observations were made entirely ou Filaria found in the abdomen, or believed
to be in the abdomen. Not suspecting this migration, and finding metamorphosis
going on in the abdomen, I may have, unwittingly, included some of the viscera of the
former in my examinations of the latter. On that occasion I traced out the metamor-
phosis to its conclusion, and entirely in what I thought, at the time, were abdominal
tissues. It is likely, therefore, that migration to the thorax is not a necessary step
indispensable for the welfare of the parasite. Eut it is certainly the usual first step for
the animal to take, and it is a fortunate one in the interests of the observer, as in tracing
the subsequent steps of the metamorphosis, the ova, which in the abdomen are so
annoying from their obscuring the field when ruptured, are not encountered. Lewis was

FILAEIA SANGUINIS I10M1NIS IN THE MOSQUITO. 373

the first to mention this migration ; until I had read his description of his experiments
on Filaria metamorphosis, I entirely overlooked this significant point.

Most of those Mlariis that do not migrate, gradually hecome granular, their outlines
hecome dim and undefined, and their movements cease. I suppose they are finally
digested.

If we examine carefully a Filaria that has just reached the thorax, we find that the
striation observed in specimens from the abdomen has disappeared ; the body is beau-
tifully transparent ; there is no sheath or sign of double outline ; oral movement still
persists, but the general body-movement has slowed down till, in some instances, the
animal is almost passive. The body is somewhat thicker and, it may be, shorter, and an
obscure undefined cloudiness can be seen in the interior (figs. 3 & 40). How far this
dissipation of the transverse striation, slowing of movement, and swelling may be owing
to endosmosis of the water that we necessarily employ in mounting the specimen, it is
difficult to say. I know that if the youn^ Filaria is soused in a large quantity of water, it
becomes so transparent that, under a low power, it is readily overlooked or even cannot
be seen. Sudden endosmosis, I believe, often ruptures and dissipates it. I therefore, in
examining the thorax of the mosquito for the Filaria at any time during the first two or
three days of the metamorphosis, employed a solution of sulphate of soda (sp. gr. 1050)
to tease up the tissues in. After application of the cover glass, the soda solution was
gradually diluted by placing a drop of pure water at one edge of the slide, while the
solution was withdrawn by a piece of filter paper placed at the opposite edge. The soda
solution shrivels, and renders irregular the outline of the Filaria, but by gradual dilution,
as described, the little animal is rendered plump and natural-looking again. I believe
it is the sudden immersion in water that is dangerous to the integrity of the Filaria.
For convenience in description, I divide the metamorphosis into six stages *. What I
have now related constitutes the first stage, viz. — ingestion by the mosquito, transverse
striation, casting of the sheath, and subsequent migration. At the conclusion of this
stage it is quite a common thing to find, with very little searching, 30 or 40 Filaria? in
the thorax, and very many more can usually be detected half concealed by careful
focussing? anions; the viscera and fragments of muscle.

The first noticeable change in the Filaria after migration is a shortening and general
thickening of the entire body, the extreme tip of the tail being alone exempted from this
process. This part retains its original dimensions for a time, while the remainder of the
animal continues to swell. Thus the caudal appendage, characteristic of the Filaria
during the greater part of its stay in the mosquito, is formed. In some specimens a
thick hyaline-looking substance seems to cover as an integument the body of the animal
from the mouth downwards, stopping short, however, almost abruptly at the root of the
tail. The tail in all cases is evidently of the same structure as the interior and mass of
the body. In others, the integument I mention is not observable, the tail seemingly

* The reader must bear in mind that this division of the metamorphosis is entirely artificial. No such thing exists
in nature. What I describe as stages, in reality overlap each other ; the graduations of development insensibly merge
one into another.

374 DE. P. MANSON ON THE METAMOEPHOSIS OF

being differentiated from the rest of the animal simply by its not having partaken in the
general swelling (figs. 4-10-38, 41, and 42).

At this stage movement still persists ; but as the body swells, motion becomes fainter
and more intermitting, till, in many examples, it finally ceases soon after migration has
been effected. There is a tendency now for the oral end to become tapered, and more
conical, one or more black points being visible at the very end.

The bodv continues to swell till it has more than trebled its original diameter, and
become shorter by one third. Thus the tail comes to look a mere appendage stuck on
one end of a sausage-shaped mass.

At first, and often all through the subsequent steps of the metamorphosis, the tail
does not correspond with the axial Hue of the body but is quite to one side. In others,
especially at a later stage, the position of the tail appears to be, and possibly is, axial,
although this effect is produced in many instances by the position the animal is viewed
in, according as the tail is to right or left, above or below ; in both the latter positions it
must appear axial, although it may really be quite to one side.

When the body has attained a thickness of about the fifteen hundredth or the two
thousandth of an inch it lies sometimes outstretched, sometimes gently bent, occasionally
curved on itself, so that mouth and tail are almost touching (fig. 42). The oral end
gives the impression that there is an orifice there, but that it is firmly pursed up.
Movement of the body is now rarely seen. Occasionally in some there is a slow general
bending and extension, and in others a peculiar shivering movement. The tail still at
times is vigorously flexed and extended. ]STo organ or structure of any definite kind can
be made out even with a high power ; only a little in advance of the tail, where the anus
is afterwards formed, sometimes an aggregation of minute cells with relatively large
nuclei can be seen (fig. 4). The rest of the body is made up of a cloudy granular
material enclosed in a very delicate integument. The latter can be made out only after
crushing of the body and partial escape of contents, or when the slide has partially dried.
In such circumstances we may sometimes see a funnel-shaped mouth into which the
delicate integument is reflected (fig. 10).

The thickening of the body, differentiation of the tail, and first indications of a
distinct mouth constitute the second stage of the metamorphosis, and usually require
two or three days for their completion.

About the third or fourth day we often find the body of the sausage-shaped Filaria,
which hitherto has shown little or no structure, filled Avith a crowd of minute shiniug
granules (fig. 11). These, I believe, are nuclei which from extreme transparency and
refracting properties, similar to that of the stroma of the animal's body, could not at au
earlier stage of development be recognized. In others again, apparently somewhat
more advanced, cells to which these nuclei belong become obscurely visible (figs. 12 to
20). These cells are exceediugly minute. Each possesses a distinct nucleus, aud
together they make up the entire mass of the body and tail inside the delicate cuticle.

About the time the cells become visible the rudiment of the future anus shows itself
as a sort of pit or vacuole a short distance in front of the tail (figs. 12, 13, 15, 39, etc.).
When viewed in profile this pit or vacuole seems infundibuliform, and the outline of

PILAEIA SANGUINIS HOMINIS IN THE MObQUJTO. 375

the margin of the body of the animal is broken at this point. Thus it would seem that
here there is a hole in the cuticle. When looked at directly from behind, or in front,
it appears round. From its first appearance, and until the last step of the metamor-
phosis is entered on, granular matter and minute nucleated cells are seen to escape
from this point. Very little pressure and sometimes even the mere immersion in water
is sufficient to bring about this granulo-cellular discharge from the anus. I do not
consider that this discharge comes from iuside an alimentary canal, but look upon it as
perianal tissue which the exterior tenuity of the walls of the rectum, and the absence
at this point of integument, permit to escape, rupture being favoured or brought about
by endosmosis of water used in mounting

Coincident with the appearance of anus and cells the mouth advances in development.
From being pursed up it seems to open, and gradually four large fleshy lips are
fashioned (figs. 16 to 30).

Then a line, at first very faint and broken, shows itself. Running in the axis of the
body for the most part, its destiny appears to be to connect mouth and anus. Around
this fine as it becomes thicker and longer, cells are seen to be arranged (figs. 14 to 25).
The line does not seem to extend in all cases quite to the anus, but terminates, as far as
one can make out, in advance of this some little distance among certain comparatively
large and prominent cells. By degrees this line, the rudiment of the alimentary canal,
becomes thicker, and the arrangement of nucleated cells around it forming the walls of
the alimentarv tube becomes verv distinct. In some instances the line itself is manifestlv
double-tubular. Probablv a fine membrane lines it. continuous at the mouth with the
skin, and is the lining-membrane of the future alimentary canal.

How this alimentary line is produced it is difficult to say. It is certainly not a
process derived from the integument dipping in at mouth or anus ; for, in many
instances in which it is visible, it is only so about the centre of the animal, and is not
traceable into either mouth or anus. My impression is. that this line is produced
something after this fashion : — After the cells I have described as filling the bodv at the
beginning of the third stase have been formed, thev arrange themselves into two sets.
One set goes to the periphery of the body to form future muscular and fibrous walls ;
the other set accumulates in single file in the centre of the bodv in a line extending
from mouth to tail. The cells thus arranged in single tile divide and subdivide longi-
tudinallv as regards the axis of the Filaria, the lines of division radiating from the centre
of the original file of cells, like the segments of an orange. This would have the effect of
making the point of convergence of the Hues of cleavage very distinct — the alimentary
line. After a time the cells swell out or increase laterally, thus opening out their hue
of convergence, which thus is converted into a tube, the alimentary canal. This process
extends no further back than the anus ; but the central line of cells can sometimes be
traced as far back as the tail, and it is principally from having seen this post-anal line of
axial cells that I infer this method of explaining the formation of the alimentary line
and canal.

■Wlide the alimentary line and canal are being formed, cells range themselves, as I
have just said, around the periphery of the Filaria, lining the skin. But between this

SECOND SERIES. — ZOOLOGY, VOL. II. 55

376 DE. P. MANSON ON THE METAMOEPHOSIS OF

tegumentary set of cells and those forming the alimentary canal there is a distinct
interval, a sort of peritoneal cavity, in which at a later stage the alimentary tube moves
freely, not being attached at any point save at the anus and mouth (fig. 25).

At the conclusion of this, the third stage of the metamorphosis, the Filaria measures
from xno" to ^0-" in length, by gro" *° sou" in breadth, or thereabouts. There is considerable
diversity both in size and shape. The mouth is wide open ; the tail is large and sickle-
shaped, and the cells of the body usually dip into it. The alimentary line runs from
mouth to anus ; and the cellular nature of the entire animal, with the exception of the
integument, is easily demonstrated. Motion is entirely suspended.

Growth hitherto has been very slow, but now, when the fourth stage commences, it
becomes rapid, the animal quickly attaining a length of from a seventieth to a fiftieth of
an inch. The walls of the alimentary canal can be distinctly traced as the walls of a tube
from the open mouth to near the anus, but not quite into structural continuity with the
latter. The cellular structure of the mass of the body is often beautifully distinct (figs.
23, 24, 25, 26). The body retreats from the tail, which becomes a mere empty integu-
mental appendage, so transparent that it can with difficulty, in many cases, be made out
(fig. 34, 35). The addition of water to the preparation causes the little animal often
to rupture after a few minutes at one or two points, a cloud of cells and granules escajfing.

In not a few instances a vacuole, similar to that which preceded the formation of the
anus, is seen at this stage at a point some little distance behind the mouth. This may
be the result of endosmosis, but possibly it is the forerunner of vulva and organs of
generation. In one instance (fig. 23) I detected a line among the cells which, from its
position, forcibly suggested that it was the rudiment of organs of generation ; but as
this was a solitary observation, I attach no great weight to it. The vacuole, in the
position I mention, is quite a common occurrence.

When the body has attained its maximum thickness the fifth stage commences. The
cells, especially in the anterior part, gradually lose their distinctness, and the mouth
inclines to purse up, while the animal, as a whole, elongates and attenuates. In some
the lengthening and thinning begin at the head and extend backwards, so that if we
find such a Filaria in the middle of this extension-process, it has a narrow anterior half
rather abruptly swelling out into a thick posterior half with the dimensions of the
previous stage (fig. 27). Such a specimen resembles in shape a hock-bottle. A very
few elongate anteriorly and posteriorly simultaneously ; and we may find in one under-
going this process a small segment of the middle of the body still thick and unextended
(fig. 28). More generally, however, the extension-process occurs simultaneously
throughout the body, the fore part being always rather in advance of the rest (figs. 29, 30).

Sometimes at this stage, just before the mouth closes up, the alimentary tube is seen
very distinctly. It moves freely in the body-cavity, accommodating itself to the gentle
movements which the anterior half of the Filaria now exhibits. The pharynx in such a
specimen is distinctly indicated by two short parallel or curved dark liues, terminating at
one end in the mouth, at the other leading into the oesophagus (fig. 44). At its other
extremity the oesophagus ends in a thick bulbous valvular-like arrangement opening
into the intestine, which in its turn may be traced as a distinct tube almost to the anus

FILAKIA SANGUINIS HOMINIS IN THE MOSQUITO. 377

(figs. 2G, 29, 30). When the mouth closes, as it does presently, all or nearly all t ace of
viscera and all trace of cells vanish. The body assumes a fibrous and very transparent
look, and little structure can be made out.

The movements just alluded to begin in the neck of the Fllaria. At first they are of a
gentle to-and-fro swaying character ; but gradually, as the whole body becomes attenuated,
they extend backwards pari passu with the attenuation, and until the entire length of the
animal becomes more or less animated. These movements do not last longer than a few
minutes, at this stage, after immersion in water, differing in this respect very much from
what happens at the next stage.

The fate of the sickle-shaped caudal appendage, which up to this time has been so
characteristic of the growing Filaria, is a little doubtful. It may be that the extreme
end of the now truncated body gradually forces its way into the remains of the tail, from
which it had before retreated, stretching it so as to have it as a closely fitting integument.
More probably the sickle is cast off in a general ecdysis, which about this stage, I believe,
occurs. Fig. 46 tends to show that this is probably what really happens ; a new skin is
shown covering the papillated tail, inside of and quite distinct from that constituting
the " sickle." In the example from which this sketch was made, on tracing the skin,
continuous with the " sickle," forwards, a point, about abreast of the anus, was reached
where it seemed to be peeling off or breaking down.

About this time at the very extremity of the truncated tail, two or three cells become
prominent, their nuclei enlarge, and the surrounding protoplasm is increased in bulk.
These cells after a time protrude from the general surface, and gradually become
converted into the circle of three or four papilla? which characterizes the Filaria at the
end of this, and during the last stage of metamorphosis (figs. 30, 31, 43, 46, 32, 33). I
am not sure if one or two of these papilla; be not the remains of the stump of the original
tail ; nor aoi I quite sure of their number, as it is usually impossible to obtain a view in
which all can be seen at once. After a time the papillae spread out like the petals of a
flower, extending considerably beyond the margin of the circumference of the body, so
that in time they become the broadest part of the animal.

The purpose of these caudal papillae is difficult to divine. Possibly they are of use in
aiding the animal in its future journey through a human host. One can understand
how by opening out on any retrograde movement they will prevent this, give a firm foot-
hold, so to speak, to the boring Filaria, and favour forward movement. They are not
oral, as I at one time supposed.

Together, the papillae, the anus, and alimentary line are very deceptive, and give the
idea that this part is really the head of the animal, the papilla? apparently surrounding a
mouth with organs of generation opening in close proximity, as is so commonly the case
in the Filarial. But I have satisfied myself that in these features we have really to do
with the tail of the animal, not the head.

During the progress of these changes the Filaria has gradually stretched from perhaps
the fortieth to the sixteenth of an inch in length ; at the same time its breadth has
decreased one half. The anterior end tapers gradually and is then abruptly rounded off,
hardly any structure being visible about it. The posterior end also tapers slightly from

55*

378 DR. P. MANSON ON THE METAMORPHOSIS OP

the anus backwards, and its very extremity is covered with the three or four papilla just
described. Faint indications of an alimentary canal are at times discernible, but further
than this no trace of internal structure can be clearly made out. The parasite has now
arrived at the sixth and last stage of its metamorphosis, and its stay in the mosquito is
about concluded.

Such specimens as I have been describing as belonging to the fifth stage are to be
found only about the end of the sixth, and beginning and middle of the seventh day,
from 130 to 156 hours after ingestion. The Filaria of the sixth stage I have only seen in
mosquitoes from 156 to 160 hours after capture, probably about seven days after feeding.
Most mosquitoes lay their eggs and die during the fifth or sixth day after feeding,
consequently it is a rare thing to find an insect of the age requisite for the maturation of
the Filaria. Probably of twenty mosquitoes one or two may attain this. I have seen
many Filarial in the fourth and fifth stages, but very few in the sixth and last.

During this stage the swaying movements that had commenced during the fifth
gradually become more active. The animal is now no longer burst or killed by immersion
in water. On the contrary, even as we look at it, the water we have just immersed it in
seems to have a vivifying effect on the parasite ; its tissues appear to become more
consolidated, its motions more active, it seems to feel that it has got into its proper
element. It rapidly acquires great muscular power, it wriggles, twists, bends, extends,
and lashes about in all directions (figs. 23, 33). As seen with an inch objective it resembles
nothing so much as the embryo F'daria in the blood looked at through a quarter-inch ;
the tail, of course, is different, but the character of the movements is exactly the same.
One peculiar motion is strikingly like one that is often seen in the blood- Filaria. Often it
pauses in the midst of its contortions, and suddenly extends itself, remaining outstretched
for a moment, its body quivering in a sort of tetanic spasm. Those who have watched
the movements of the Filaria in the blood will recognize what I mean. In other examples
the wriggling movement is not seen, the activity of the animal being expended in frantic
rushes forwards and backwards, and in every direction. I presume this description of
movement belongs to a slightly later stage than does the wriggling motion ; probably it-
is the motion suited to the time when the mosquito has died and fallen into water, and
the parasite has obtained its liberty.

Frequently, during the contortions of the animal, it turns its mouth towards the
observer, and seems to be endeavouring to uplift the cover-glass. During this movement
we can see that the mouth is pursed up into a cone, the lips being firmly approximated ;
and, around the spot where the mouth must be, a number of exceedingly minute horny-
like papillae are arranged. This may be the boring-apparatus. But the motions of any
animal I have examined at this stage have been too incessant to permit a satisfactory
view, and these boring-papillae may be, to some extent, matters of imagination. So
incessant are the movements that it is quite impossible to obtain satisfactory measure-
ments. After watching one for over an hour, and despairing at length of getting it to
slow down, I killed it by instilling a little osmic-acid solution below the cover-glass. It
then gradually slowed down, stretched out, and died. I found it to measure -fe" x &W-
In a good light this particular parasite was perfectly visible to the naked eye.

FILAEIA SANGUINIS HOMINES IN THE MOSQUITO. 379

I believe the Filaria, after it has attained this final stage, has hut a very short time
longer to pass in the mosquito. I am inclined to think it preys on the tissues of the
insect, and thereby contributes to its death. For, in examining mosquitoes found dead
on the surface of the water, and which I know could have died only a very short time
before, I have been struck with a singular absence of viscera and muscular tissue — the
thorax seemed but a hollow shell. Again, some mosquitoes, even after death, cling or
stick to the sides of the vessel containing them ; in such specimens, also, I have remarked
an absence of viscera and muscular tissue. My impression is, that these have been con-
sumed by the Filaria, that the death of the insect had been thereby hurried, and that
when this occurred the parasite bored its way out of the body, and thence sped to the
water. I have never found a Filaria in the last stage of the metamorphosis in a dead
mosquito ; often, however, I have found them at all stages but this, even up to the time
when they measure the fortieth or the thirtieth of an inch in length. The final stretching
process must be rapidly effected, lasting probably an hour or two ; and, as during it the
Filaria becomes more than double in size, there must be expenditure of the tissues of
the intermediary host to provide pabulum for this rapid growth. Hence, in some
measure, it comes about that one rarely catches the Filaria in the last stage of growth,
and that the dead mosquito has an empty thorax.

Such, I believe, is a true description of the life of Filaria sanguinis hominis in the
mosquito. Some of my interpretations may be wrong, but all the principal facts I have
stated are true. I have verified them over and over again.

I can understand how, if fig. 1 and fig. 34 are compared, some doubts may arise as to
their connexion. The contrast between fig. 1 and fig. 23, or between fig. 23 and fig. 33,
is great — so great that unless one had carefully traced the connecting links step by step,
one would positively declare them organisms of entirely different species. Over and
over again, when working at the Filaria of the fourth or fifth day, I have hesitated to
believe it was really the outcome of the Filaria I had seen in the abdomen of the
mosquito an hour or two after feeding; and I have turned to mosquitoes of the first,
second, and third days to reassure myself of the migration to the thorax, the differen-
tiation of the tail, the swelling of the body, the formation of the mouth, anus, and
alimentary lines before I was at my ease as to the genesis of the particular organism
that I was specially engaged in studying. But now, to my mind, there does not exist the
slightest doubt in the matter.

Besides being able to trace the gradation in mosquitoes of different ages, we often
encounter specimens of the Filaria, at all stages of development from fig. 3 to fig. 14, in
one and the same mosquito. Sometimes in an insect in which a Filaria like fig. 14 is
found, many others, at all stages of development, from fig. 14 to fig. 24, may be encoun-
tered. And so with the latest stages : fig. 23 may be found in a slide in which fig. 33 is
moving. This gradation is very striking, and clearly indicates the connection between
the Filaria ingested by the mosquito and the Filaria about to leave it.

The most difficult step to follow, and the one over which Lewis apparently hesitated,
is that represented from figs. 4-12. If, however, insects of from twenty-four to forty-
eight hours after feeding are studied, transition-forms are found in abundance. Eigs. 40,

380 DR. P. MANSON ON THE METAMORPHOSIS OF

41, 42 are from the same insect. I have seen and sketched specimens resembling figs. 3,
8, and 12 in the same field of the microscope, and lying across each other.

In the little caudal appendage that characterizes the Filaria from the second to the
fifth stage we have a strong argument for believing we are concerned with the same
species during all these stages. It ceases, however, to be distinctive towards the end of
the fifth and during the sixth stage; but here a curious circumstance comes to our
assistance in establishing identity. The thick cellular body of the fourth and commencing
fifth stages is motionless ; but as the anterior part fines down and stretches out, it at the
same time acquires movement. A difference also obtains between these two parts a&
regards their capacity for being stained with gentian violet. The posterior and cellular
part takes the stain with difficulty, or not at all ; whereas the anterior fibrous and active
part takes the stain rapidly and deeply. Now we sometimes find a Filaria with a long
stretch of the fore part of the body in thickness and general appearance like that of fig. 31,
and in very active movement, dragging after it a limp and seemingly paralysed posterior and
cellular part, rupturing readily on immersion in water, and showing all the characteristics
of the cellular stage. Further, on applying the stain to such a specimen, the active fore
part is deeply affected, whereas the limp hind part is but slightly tinged * . Such a
specimen bridges over the step from figs. 23 to 33. Catching specimens such as fig. 27
and fig. 28 in the act, so to speak, of stretching, distinctly establishes that the Filaria oi
the sixth stage is but an advanced form of the Filaria of the fourth, and, by a similar j
process of reasoning, of the first, and of the Filaria in the human blood.

I can imagine that some may found an objection to the theory of the action of the
mosquito on the fact that, of the hundreds of Filaria; entering a certain mosquito, only
a dozen or so will attain the more advanced stages of development ; and that, of those
that do thus advance in development, most perish in consequence of the death of the
insect before the Filaria is equipped for independent life. It is true that the great
majority cease to develop at a very early stage ; others, after advancing a little, seem to
die and wither. Fig. 36 is from such a specimen. Perhaps only certain tissues or
organs of the mosquito favour complete development, and in their blind wandering few
of the Filarice reach these. But these facts are no argument against the theory that the
mosquito is the intermediary host of the Filaria : indeed, on the contrary, they are in
entire consonance with what Ave know of the life-histories of other parasites. Of the
millions that leave the parent, but one or two ever reach maturity. The journey from
the uterus to the spot where maturity is attained is beset with dangers, the measure of
these dangers being the prodigious numbers that start on the journey.

* Strange to say, the fully developed Filaria, as that represented in figs. 32, 33, does not take the stain well. It
seems to me that the age of the integument has something to do with the facility with which the stain is ahsorhed.
Specimens, as at tigs. 22, 23, 25, do not, stain well, I believe the reason being that the integument in these is old and,
presumably, dense. Hut the anterior half of fig. 27 would stain well: it is the growing half, and, I think, has just
undergone ecdysis, and the new skin is easily permeated by the dye. Its posterior half, however, will not stain at
all, or but very slightly and unevenly ; the old skin is not yet cast, and is dense and impervious. The new skin,
after a time, as it ages, also becomes dense, so that by the time the animal has grown to the size of the original of
fig. 33, it is impervious, and will hardly allow the passage of the gentian violet.

FILARIA SANGUINIS HOMINIS IN THE MOSQUITO. 381

Frequently during my study of the metamorphosis in the mosquito, it has occurred to
me that the later stages of what I describe as the Mlaria of man in the mosquito, may
really be stages in the development of a purely insect parasite like Gorditis or Mermis,
',and not the Mlaria sanguinis hominis at all. Perhaps, I conjectured, this animal may
escape from the body of the mosquito into water, where it may acquire organs of genera-
tion and breed, and its young may afterwards enter the body of the mosquito larvae, and
so complete the cycle. But if this is the case, there must be a time in the histories of the
human Mlaria and this hypothetical animal when it is impossible to discriminate between
them. I have never been able to find a break in the chain of development.

Moreover, if the mosquito is liable to be infested by a nematode parasite peculiarly its
own, and if the developmental forms I have been describing belong to such a parasite,
then we ought to find them in mosquitoes that had fed on non-filariated, as frequently
as we find them in mosquitoes fed on filariated blood. If one filarial form could be
tbund in a mosquito fed only on non-filariated blood then my theory about the role of
this insect is in danger of falling to the ground.

This is a point I gave some attention to. Simultaneously with observations on
mosquitoes fed on the blood of my filarious subject, I carried on others on mosquitoes
saught in the mosquito-nets of coolies and servants living in my house. The result was
pery striking. I carelessly assumed that my servants were free from Mlaria, and did not
3xamine their blood, but trusted to the mosquito to pick out for me those who might be
Eilarious.

Two mosquitoes were brought to me by coolie No. 1. One of these I examined at
Dnce — no Mlaria ; the second I kept till the fourth day before examining it — no Mlaria.

Coolie No. 2 brought two mosquitoes. I examined one on the first day but found no
Filaria in the blood in the abdomen. The second I kept till the third day. I examined
ts thorax very carefully, and then had the mortification of finding a sausage-shaped body
exactly like that represented at fig. 12. There could be no doubt about its nature. I
;hought my theory was upset, and all my work useless. At 9 p.m., however, I examined
olood drawn from this man's finger, and had the satisfaction of finding in it plenty of
Filarim ; subsequently he contributed Mlaria-chavgeA mosquitoes for my investigation
»f the metamorphosis. The mosquitoes he brought me were often devoid of Filarice ; he
jvas a lazy man, slept much during the day, and in a very dark room such as mosquitoes
iffect even during the day, hence many of the insects he found in his net were charged
vith blood drawn at a time when the Filaria is not in the general circulation.

Coolie No. 3 brought on one occasion two, and on another occasion five mosquitoes.
Their examination gave this result : —

Mosquito examined on 2nd day No Filaria.

„ „ on 3rd day „

„ „ at 74 hours ,,

,, at UG hours One sausage-shaped Mlaria.

n „ Two Filarice.

„ „ Many Filarice.

„ Several Filaria.

Again my theory seemed doomed to collapse. But I examined the man's finger-

382 DR. P. MANSON ON THE METAMORPHOSIS OF

blood in the evening with the same results as in coolie No. 2 ; it contained plenty of
Filarice.

Coolie No. 4 brought six mosquitoes with this result, that three of them when ex-
amined ninety-six hours after capture were found to contain Filar ice, and three of them
yielded no parasitic form whatever. This man's blood was examined at 9 p.m., and, like
that of coolies No. 2 and 3, was found to contain many Filarice. These coolies all came
from Hoorcoah, a district to the north of Amoy, in which a very large proportion of the
inhabitants are filariated.

It was now evident I must examine mosquitoes from some less questionable feeding-
ground before I ventured to draw any conclusion from this endeavour to prove a negative.
I had a boy of about 15 years of age for a servant ; he was healthy and his blood did not
contain Filarice. I got this lad to bring me his mosquitoes, with this result : — Eight
mosquitoes examined sixty-four hours after capture — no Filarice found ; one mosquito
examined eighty hours after capture — no Filarice ; one mosquito examined eighty-eight
hours after capture — no Filarice ; five mosquitoes examined one hundred and four hours
after capture — no Filarice. I examined a few mosquitoes fed on other non-filarious
individuals, and with the same negative results.

Many years ago, in searching for the intermediary host of the Filaria immitis, I ex-
amined many tens, if not hundreds, of mosquitoes that had fed on dogs' blood ; but in no
instance did I remark, although I was on the outlook for such, an advanced form of
filarial development.

Lewis also did not find sausage-shaped nematodes in any mosquitoes but in those that
came from a certain room where it was afterwards discovered that a filariated man slept.
He examined with a negative result many mosquitoes that had preyed on the blood of
dogs or birds.

Similarly Sonsino found Filaria? in about half his mosquitoes ; they were captured in
a bed where two men slept, one of whom had Filarice in his blood, whereas the other was
not thus infested.

The inference from these facts is obvious. It is only mosquitoes that have fed on the
blood of filariated men that contain the developmental forms I have described.

In the foregoing I think I have conclusively shown that the mosquito fed on filariated
human blood ingests the embryo Filaria sanguinis hominis, that while in the mosquito
the Filaria undergoes great developmental changes, and that it finally quits this insect as I
a large and powerful animal equipped for an independent life. I have also shown that the
mosquito fed on non-filaviated blood exhibits, when dissected, no such parasitic forms. I
therefore infer that the parasitic forms I meet with in the mosquito fed on filariated blood
are really advanced forms of the embryo Filaria ingested with the blood, and that the
mosquito is the proper intermediary host of the Filaria. I do not see how this inference
can be avoided. It would be much more satisfactory, and an easier method of demon-
stration, could we watch an individual Filaria in its progress from the circulation down
the proboscis of the mosquito into its stomach, watch it wriggling there, case its skin,
and, changing its method of movement, worm its way through the tissues into the
thorax, become passive there among the muscles while it acquired size, an alimentary

FILARIA SANGUINIS HOMINIS IN THE MOSQUITO. 383

canal and other organs, and enhanced muscular power ; then eat or bore its way through
the integuments of the dying mosquito, and finally obtain liberty in the water into which
the insect has fallen. It is impossible to do this ; but, by comparison of a large number
of dissections, we can follow the history of the animal in the mosquito almost as perfectly
as if we had watched in transparent tissues the progress of a single individual from the
human body to the water.

The experimentum cruris of this theory, as I have already said, I have not had the
hardihood to attempt. But from what I have written, any one anxious to make it will
have no difficulty in gathering what is likely to be the most successful method. Were I
to attempt it I would proceed in this fashion : — I would feed my mosquitoes on a filarious
subject, I would collect them every morning, giving each a bottle to itself. Those that
survived to the afternoon of the seventh day I would transfer to test-tubes ; these I would
invert over a watch-glass, containing a little water. When the insect died I would allow
it to remain on the water a few hours and then remove it. The water it had fallen into,
and which probably now contains the Filar ia, I would transfer to a stock-bottle contain-
ing water. This process I would repeat for several days. I would then administer
portions of the contents of the stock-bottle to the subject of the experiment. I would
continue this for a month, every day adding fresh mosquito-water to the stock-bottle,
and every day administering a draught of its contents. After a time I would commence
the examination at night of the finger-blood. I am quite satisfied as to what would be
the result.

EXPLANATION OF PLATE XXXIX.

[All the figures with the exception of figs. 10, 34, 35, 37, 43, 44, 45, and 46, are magnified about

188 diameters.]

Fig. 1. The embryo Filaria sanguinis hominis as it appears in the blood, or lymph, or in the abdomen of
the mosquito immediately after ingestion, ^V" x xobo"-

Fig. 2. A Filaria about an hour after ingestion by the mosquito. The sheath has been cast, transverse
striation and oval pouting are very distinct, and the animal is indulging in the snake-like
wriggling by which it moves from the abdomen to the thoracic viscera. In the mosquito from
which this specimen was obtained many Filaria were found in the newly ingested blood in the
insect's abdomen. All were active and transversely striated. Most had cast the sheath. In
one the sheath was lying at some distance, in another it trailed after the animal, while in a
third it lay across it. Oral pouting was distinct in all ; but no double outline, or further struc-
ture, could be detected in any of them. In the same insect two Filaria were found in the
thorax ; they were structureless, without sheath, somewhat swollen from endosmosis, an obscure
convoluted granular-looking condensation occupying most of the body.

Fig. 3. From the thorax twelve hours after ingestion, y^/x g0loo"-

The abdomen of the insect was half filled with blood, in which moved many active, trans-
versely striated, pouting Filaria. They had no sheath. In the thorax many were found.

SECOND SERIES. — ZOOLOGY, VOL. II. 56

381 DR. P. MANSON ON THE METAMORPHOSIS OF

Tliey showed no double outline, transverse striatiou, or sheath ; hut oval pouting and some
general body-movement still persisted. As yet tail-differentiation was not discernible. Beyond
a few undefined specks the body was homogeneous. Perhaps the mouth possessed four lips.
Slide prepared with sulphate-of-soda solution gradually washed away with pure water.
Fig. 4. From the thorax 25 hours after feeding. The abdomen of the mosquito still contained fluid, or
semifluid blood, about two thirds full. In this eight passive Filarice were found ; they were
granular and evidently dead or dying. Over two dozen Filarice were found in the thorax, most
of them active, with differentiation of the tail commencing. That represented measured

loir A 2Boo •
Figs. 5, 6. From the thorax 35 hours after ingestion. The abdomen of the insect was about two thirds filled

with blood, in which 12 or 13 undefined, granular, dead, and fading Filarice were found. One of
them, however, which evidently had been about the periphery of the clot, was well defined, faintly
striated, plump, moving freely, and had a tail slightly differentiated. Many were found in the
thorax ; all of these were passive and had distinctly differentiated caudal appendages. No
organs seen; body-contents granular. The body appears to be expanding, the enlargements
being most apparent towards the tail end. One measured xio"x ^oW'j another x4o"x Woo"-
One or two exhibited slight movement, cither slow bending and extension, or, at loug intervals,
sudden jerking extension.

Fig. 7. From the thorax, 39 hours. Large number of Filarice in the thorax. For the most the tail was
well differentiated. The body of the parasite had become thicker and shorter. In some, on
washing away the sulphate of soda solution, a double outline could be made out. Contents
cloudy and granular. Jerking movements frequent. A very few of the Filarice were shaped
and moved as in mosquitoes examined a short time after feeding ; that is, they were long
slender, active, with a tapered tail. Between these and the distinctly tailed animal transition-
forms were abundant, so that there could be no doubt about the connexion. Specimen repre-
sented measured xis" x or,\M)".

Figs. 8-10. From the thorax of mosquito 46 hours after capture. Many Filarice were found in this
insect at different stages of development, from the long, slender, rather active Filaria with but
slightly differentiated tail, to the fat tailed sausage-form with granular contents and intermittingly
wagging tail. Fig. 10 shows the chitinous integument, dipping into the rudimentary mouth;
it is drawn to a larger scale. In one field of the microscope three Filarice lay across each other;
one had hardly made any progress in development, another was still slender, but the tail was
differentiated, while the third was thick, plump, and sausage-shaped.

Figs. 11, 12. From the thorax 56 hours after feeding.

The insect was found dead at 2 p.m., but was seen alive the same morning. Many live
Filarice found in this insect, exhibiting the extension of the body and flick of tail movement.
Fig. 11 showed some body-extension and a remarkable vibration of the whole of the posterior
part of the body. The contents were in marked contrast to those of No. 12, being a structure-
less homogeneous matrix interspersed with innumerable, well-defined, shining granules ; it
measured X2s"x rnJo''* Fig. 1~ measured X4s''x x^Su"' ^ possessed no body-movement, but
only the tail-flick. Its body was made up of distinct cells ; one minute, glistening, nucleated,
cell-like body was a prominent feature at a point where the anus might be. Many more
Filarice like these two types in the slide.

Figs. 13, 14. From the thorax 72 hours after capture. Many Filarice much broader, and somewhat
longer, than those from the mosquito of figs. 11, 12, which belonged to the same batch of insects,
and examined the day before. The tail was not seen to move, and no movement of body
could be detected. Obscure cellular arrangement of the body. In all specimens a crowd of
granular matter escaped from the anus. Fig. 14 measured xio"xToo" anteriorly, and -s\-5n

FILAEIA SANGUINIS HOMINIS IN THE MOSQUITO. 385

posteriorly ; it was slightly constricted in the middle. After the slide had been mounted some
time, a shrinking of the contents showed the integument of the parasites very distinctly, all
exhibiting a marked doable outline.

fig. 15. From the thorax of a mosquito the date of whose feeding could not be fixed, probably about 80
hours before examination. In its thorax two Filarice were found, one partly crushed, the other
perfect— as represented. Measurements T]-0" x ^Vo" anteriorly, and 8^ff" posteriorly. It ex-
hibited occasionally spasmodic jerks. At the head a distinct double outline. An alimentary
line ran backwards some distance, becoming convoluted posteriorly, and gradually becoming
less distinct till it was lost at the middle third. Anus distinct, open, and emitting a few
granules. Tail distinct. Contents of the hody obscurely cellular, lobulated, granular.

?i°-. 16. From thorax 97| hours after capture. A considerable number of Filarice found resembling
sketch. Body smooth in outline, plump, and healthy-looking. The mouth was four-lipped,
and a pharynx could be made out ; but no other trace of an alimentary canal unless in one
specimen in which a funnel-shaped anus was apparent in the usual situation. The tail was
delicate, but moved distinctly. There seemed to be a condensation of tissue near the surface of
the hody, especially about the waist. There was a shoulder and an anal bulging. Movement
of the lips was observed. Measurements T£0"x 1200"-

Fig. 17. From thorax 97 hours after capture; about a dozen sausage-shaped Filarice found. Four lips
readily made out; alimentary line running from mouth to anus. Cellular matter escaping
from anus. Content.-; of body cellulo-granular, and very obscure. Tail small. Specimen

measured i\{,"x -7J,„"-

t Figs. 18, 19. From the thorax of a Filaria found dead (but recently living) 96 hours after capture.
' Alimentary line very distinct. Body celluio-granular. Four lips rather pursed up. One
measured Tf0" X 7}/. Many similar in this insect.
Fig 20 From thorax 144 hours after capture. In this particular insect eight such Filaria were found.
Mouth, alimentary canal, and anus, well marked in all. Tad was very faint, and it could
easily 'be ascertained that at this stage it was nearly entirely integumented. The longest
measured fa" x -gW • In it the cellular structure of the body was very apparent posteriorly ;
towards the°head the cell-outlines could not be made out, the tissues looking granular or homo-
geneous. The mouth was pursed up. No movement remarked; granular escape from anus
considerable. Alimentary line very thick, giving the impression that it was filled with some
dark substance.

; Figs 21-23 From the thorax of a mosquito 128 hours after capture. In this insect about a dozen
advanced Filarice were found. They were nearly all straight, extended, with truncated posterior
ends tipped with a minute, delicate, tegumental tad. Mouth distinctly four-lipped. Shoulder
and anal bulging, and waist, inmost. Alimentary canal traceable from the mouth to anus in
some in others gradually losing itself about the middle of the body. Round this line nucleated
cells have ranged themselves. In fig. 21 a prolapsus of part of the intestine has occurred.
The body is filled with minute, clear, nucleated cells, in all the specimens. In 23, in addition
to the alimentary line, traces of what may be commencing organs of generation were seen; a
line could be followed from a point a little posterior to the head meandering backwards for some
little distance among the cells. Where this line opened on tbe surface a dimly indicated infun-
dibulum, which maybe vulva and vagina, was indicated. Outline of all distinctly double.
Fig 23 measured fa" X 5TjTj "j the others about fa" x ty and fa" x *fo".
Fig 24 From thorax 145 hours after capture. Eight Filarice in thorax far advanced in development.
' The mouth and tail very evident. In some the tail was entirely integumental ; m others a
papilla from the body projected slightly into it-fig. 34; whilst in others the integument was
quite filled up with the same material that made up the bulk of the body-fig. 35. When first

386 DR. P. MANSON ON THE METAMORPHOSIS OF

seen, i.e. soon after the body of the insect was immersed and broken np in water, part of the
alimentary canal was visible as a distinct tube, and the body was for the most part veiy clear.
After a time the alimentary canal became obscured by the cellules becoming granular and con-
fused. Granular matter escaped from the anus. The cellules were smaller than those in
figs. 21, 22, & 23. Several parasites were crushed under the cover-glass ; in these chitinous
integument was readily seen. Measurement y's" x ■g^g".
Fig. 25. From the thorax of a mosquito found dead on the morning of the seventh day after capture.
The arrangement of cells round the alimentary canal, and on the inside of integument, very
easily made out. The cells were plump and nucleated. In addition to the peripheral layer of
cells the walls of the body appeared to be strengthened by a sort of thick corium, to the inside
of which the cells were attached. There was a distinct space, a sort of peritoneal cavity,
between the cortical and intestinal cells. In one specimen the head moved slowly at intervals.
Mouth in all pursed up.

Kg. 26. From thorax. Mosquito found dead 144 hours after capture, seen alive 8 hours before. Fight
such Fi/ariee found. Two of them had slight movement; the largest measured -g^-" Xyj5";
another ^5" x y]5. The sickle-shaped tail was purely integumental, very faint and fine. The
body in some projected into this as a papilla ; in others it was abruptly truncated. Iu most
the anus was a large bag with well-defined wall, and from it granules escaped. Alimentary
canal traced as a tube two thirds down the body, but union with rectum not distinctly discerned.
Mouth widely open. Cellular structure of body obscure. Many shining granules in the wall
of the alimentary canal.

Fig. 27. From thorax of mosquito found dead and adhering to damp side of bottle, 132 hours after
capture. Altogether there were about six Fi/ariee in the insect's thorax, and all in an advanced
state of development ; mouth, anus, alimentary canal, distinct ; body cellulo-granular. Mouth
in every case pursed up so that the lips were with difficulty defined. The specimen represented
was more advanced than the others. The anterior part of body had thinned down and
lengthened, while the posterior part remained full size, giving the animal the hock-bottle
appearance. The cephalic end very transparent, alimentary tube in it barely visible. Mouth
pursed up, and no lips discernible. Cephalic end exhibited slow swaying movement when fresh.
Measurements ^"xjo'jo" anteriorly, and 5J-5" posteriorly.

Fig. 28. From mosquito 135 hours after capture. Many advanced Filarice found in this insect. Some
of them showed thinning and commencing extension of cephalic end. The specimen repre-
sented had grown at both extremities, particularly at the cephalic ; but a portion of the middle
of the animal had not yet become extended, hence its peculiar appearance. The entire animal
measured ^" in length ; the neck was xoV>o"> tue posterior part gl 0", and the bulging centre
-1,," in diameter. Lips distinct ; body cellulo-granular; tail, a long delicate sickle. Some
motion exhibited.

Fig. 29. From a mosquito 144 hours after capture. Many advanced Filarice in thorax, some of them
undergoing stretching. One measured t^" x ^s" ; another j}() x j\t)". Oesophagus could be
made out as a thick-walled tube in most ; in some it terminated in a bulbous valvular-looking
arrangement, from which sprang the intestine. Mouth distinctly lipped ; a pharynx seen in
some. In some no tail could be detected, in others a very transparent sickle was seen. Body
cellulo-granular ; no movements.

Fig. 30. From mosquito 152 hours after capture. About 15 large Filarice were found in the thorax.
The largest is represented, it measured ^q" by ^\5" posteriorly, 7^/' anteriorly. A number
were found crowded together in one corner of the slide at various stages of extension from the
TJ0" up to the r-y in length. Cellular structure of the body in some most apparent. Oeso-
phagus in some particularly distinct, the single layer of nucleated cells constituting the wall

FILARIA SANGUINIS HOMINIS IN THE MOSQUITO. 387

showing out well. Mouth open, four-lipped, with signs of papillae on skin (fig. .37). In some
no trace of tail, in others a large beautifully transparent sickle into which the body dipped a
stumpy papilla. Rectum very evident. Granules in some escaping. A little, but very little,
movement in some heads. No trace discoverable of organs of generation.

Fig. 31. From the thorax of a mosquito 156^ hours after capture. The ova had been deposited. In
this insect Filarice were found at almost all stages of development, from large " sausages " to
almost the Filaria of the sixth stage. One of the latter measured §■$" x yooo", and had three
distinct papillae on the tail. In this instance motion was most active in the tail, but the head
also was animated ; a prolapsus of intestine took place about the middle of the body. In
another, about ^V' long, the alimentary canal was very distinct, but rapidly became granular;
hernia occurred in two places, and the tail and head retreated from their integuments ; in the
case of the head it left unoccupied about yoVo" °f the integument, and in this the involution to
line the pharynx was very distinct. In some specimens the transparent sickle-shaped tail still
remained.

Figs. 32, 33. The former is somewhat diagrammatic ; the latter was constructed from observation of two
full-grown Filarice at the sixth stage. The following are my notes of examination of the par-
ticular insects in which the originals were found : —

(a) A large brown mosquito, 158i hours after capture, with ova still undeposited. Many
Filarice were found in the thorax, in the fifth stage, measuring from g1,/' to £$", or thereabouts.
One Filaria had attained a more advanced stage ; it measured -^" x ^Vs"- It possessed great
activity, for it was in incessant motion, wriggling, coiling, and uncoiling. One movement was
peculiar ; it slowly, slightly approximated one end of the body to the other, and then suddenly
extended itself. The motion was too quick for the eye to follow. The viscera were very
difficult to make out. There were papillae at one end of the body, the other end was tapered,
and then abruptly rounded off. No trace of organs of generation visible. No trace of alimen-
tary canal, or anus, could be made out, all details being obscured either by extreme opacity or
extreme transparency of the body. Unlike Filarice at an earlier stage, this individual seemed
quite at home in the water ; the longer it remained in it the more active it became, and there
was no bulging of the body or escape of contents. The other Filarice found in this mosquito,
after moving languidly for a time, gradually became less active, and finally passive. Their
bodies at one or two places would bulge and then rupture, intestine protruding. In these
Filarice details of structure were easily made out. The large Filaria was visible to the
naked eye.

(b) A large brown inosquito, 159i hours after capture, eggs still undeposited. The thorax
contained about 8 Filarice, " sausages " for the most part at an advanced stage. One measured
^y ; another measured ^" x %i3". The fore part of the body of this specimen was in con-
stant, active, swaying motion ; and the posterior half of the body apparently in the advanced
fifth stage, i. e. it looked more cellular, was affected by endosmosis, and possessed little motion.
The fore part had a firm, fibrous look, and the alimentary canal could sometimes be seen moving
inside the muscular and fibrous walls, accommodating itself to every movement. The mouth
was firmly pursed up, and the outline of the lips was quite indistinguishable. As it turned at
times towards the eye, the very centre seemed as if pointed to a spine, a ring of four or more
little spines, or papillae, surrounding it ; this is to some extent imagination, the perpetual move-
ment preventing reliable observation. On the extremity of the tail were two, perhaps three,
large, rounded papillae ; close to these, about two diameters of the body in advance, the anus
and rectum could be distinctly made out. A hernia of the intestine occurred after this animal
had been under observation about an hour. The hernia consisted entirely of intestine, no trace
of uterine tubes or testicle. The movement of the head end continued for an hour and a half,

SECOND SERIES. — ZOOLOGY, VOL. II. 57

388 METAMOKPHOSIS OF FILARIA SANGUINIS HOMINIS IN MOSQUITO.

apparently brisker at the end of this time than at the beginning ; it was only suspended when
the cover-glass was removed, and the slide dried.

Figs. 34, 35. Show with tig. 24 the process of withdrawal of the body from the tail.

Fig. 36. Probably a dead Filaria with irregular bulgings, vacuoles, and signs of degeneration.

Fig. 37. Head of fig. 29 more highly magnified, and drawn at a slightly later hour, when the body on
dying had somewhat retreated from the integument.

Figs. 38 & 39. From the same mosquito, showing differences in the rate of progression of meta-
morphosis.

Figs. 40, 41 & 42. Also all from the same insect.

Fig. 43. Caudal papillae developing.

Fig. 44. Pharyngeal lines very distinct, from a Filaria in the fifth stage.

Fig. 45. Showing a sort of horny thickening of the tips of the lips at the fourth and early fifth stages.

Fig. 46. The sickle-shaped integumental caudal appendage not yet shed; caudal papilla? developing,
and possessing a delicate skin separating them from the original integument of the earlier

[389]

XIV. On the Compound Vision and the Morphology of the Eye in Insects. By
B. Thompson Lowne, F.R.C.S., E.L.S., Lecturer on Physiology , Middlesex Hospital,
formerly Arris and Gale Lecturer, Royal College of Surgeons.

(Plates XL.-XLIII.)

Eead 7th February, 1884.

J_HE manner in which the compound eye of arthropods subserves the function of
vision has been an undetermined problem since Johannes Midler enunciated his well-
known theory of mosaic vision in 1826 *.

The views of naturalists on this subject, if we except the extremely improbable and
purely hypothetical view recently propounded by Exner f , may be grouped under two
heads : — some have supported Midler's hypothesis, or a modification of it % ; whilst others
have followed R. "Wagner, and held the view more commonly attributed to Gottsche §.

It is well known that Midler supposed that each facet produces a single visual im-
pression and that the whole visual field consists of a mosaic of such impressions ; many
of his followers have modified this view, by supposing that a small number of visual impres-
sions are originated by each facet, the order of which is neither reversed nor inverted by
the dioptric apparatus.

Midler's hypothesis was suggested to its learned author by the radial arrangement of
the parts of the compound eye about a hemispherical or nearly hemispherical retina,
and by the difficulty of conceiving a retina capable of correcting a mosaic of reversed
and inverted images, the order of which is neither reversed nor inverted.

R. Wagner || was the first to throw doubt upon Midler's view : he mistook the capsule
of the crystalline cone for a retina, an error which was also committed by Ruete If and
Dor **.

R. Wagner and his followers regard the compound eye as an aggregation of simple
eyes, the dioptric structures of each producing an image on a distinct retina, in the
same manner as the dioptric structures of the vertebrate eye, so that the whole visual
field is a mosaic qf reversed and inverted images, the order of which is neither reversed nor
in verted.

The most important paper on this subject, after Midler's work, was undoubtedly a
short but well-known contribution to Midler's Archives by Gottsche in 1852. Leeuwen-

* J. MiUler, ' Zur verglcichenden Physiol, des Gesichtssinnes des Menschen und der Thiere ' &e.
t Sigm. Exner. Biologisches CentraJblatt, Jahrg. i. p. 272, and Pop. Science Eeview, 1881, p. 337.
J Holmholtz, Du Bois Eeymond, J. Notthaft, &o. § Will. Zenker, &c.

II Wiegniann's Archiv, 1835. Ed. i. p. 372.

1" Gratulationsschrift der mod. Fac. zu Leipzig zu C. G. Cams 50-jahr. 1861.
** Archiv d. Sciences Phys. et Natur. 1861.
SECOND SEKIES. — ZOOLOGY, VOL. II. 58

390 MR. B. THOMPSON LOWNE ON THE COMPOUND VISION

hoek had long before observed the images produced by the individual facets. Gottsche
observed and drew attention to these : he, I believe, was the first who investigated the
structure of the great rods (Sehstabchen), and suggested that the highly refractive axial
threads, which he discovered in their interior, in some way transmit the images formed
by the lenses to recipient nervous structures beneath.

I can hardly imagine that the earlier writers intended to convey the idea that an image
could be transmitted, as Dr. Grenadier facetiously suggests *, " as a message is trans-
mitted by a telegraph wire," but suspect that they meant that rays of light from certain
definite points in the image are so transmitted. There is, however, considerable
ambiguity about Gottsche's paper on this and other points, although he apparently
correctly indicated the position of the true recipient layer, beneath the great rods,
without actually observing it.

Gottsche's paper appears to have been completely misunderstood by all his critics,
who seem to have imported into it a false analogy between the image in the compound
eye, and that in the vertebrate eye : the image is thus supposed to fall upon the retina.
If Gottsche had intended to convey such an idea his view would have been, as it is
generally supposed to be, in direct antagonism to Midler's hypothesis; yet Miiller
gave a kind of adherence to Gottsche's view, holding it to be consistent with and supple-
mental to his own.

It appears to me that the whole tenor of Gottsche's paper is an attempt to modify, not
to destroy, Midler's theory, although both he and Miiller, in the note which accompanies
this paper, seem to have forgotten the difficulty which arises from the inversion of the
subcorneal image. And further, Gottsche's retina is not the retina of Wagner, Ruete,
and Dor. Soon after the publication of this paper Leydig t drew attention to the con-
tinuity of the axial structures of the great rods and the crystalline cones ; and supposed the
cones to be the terminal organs of the optic nerves. His views on the connexion of the
cones and the nerve-centre are not easy to comprehend, as he appears to deny the truth
of Midler's hypothesis, at the same time that his observations apparently support it.

ClaparedeJ then pointed out the apparent continuity of the cone and the corneal
facet, in Typhis, and found that the cornea, cone, and axis of the great rod, in some
insects, consist of the same material ; he asks, if it is not possible that these are all con-
cerned in transmitting the image to nervous structures beneath them.

In criticizing the views of Gottsche, Claparede, and Leydig, it appears to me that it
should constantly be borne in mind, that they worked at a time when the close relation
between physical and vital phenomena was less completely understood than at present,
and that they consequently, perhaps unconsciously, ascribed hypothetical vital properties
to structures, which will not bear rigorous investigation.

Max Schultze § first investigated the optical relations of the great rods, and concluded
that they are not the terminal organs of the optic nerve, after a most laborious exami-
nation of their structure. He also rejected Mailer's hypothesis as optically untenable.

* Untersuchungen iiber das Sehorgan der Arthropoden, p. 10. Gottiugeri, 1879.

t Miiller's Arohiv, 1855', p. 406. X Zeitschrift fur wissensch. Zool. Bd. x.

§ Schultze, Arohiv, Band ii. p. 404.

AND THE MORPHOLO&Y OF THE EYE IN INSECTS.

391

Boll and Grenadier have, however, more recently adopted a modification of Midler's
hypothesis, based upon a view held by Max Schultze. Gottsche described the very
remarkable structures, at the inner extremity of the great rods of the lobster, under the
term " doppel-jPyramiden "*, better known as the spindles. Max Schultze regarded the
spindle as the true nerve terminal; it is the " Hetinula" of Dr. Grenadier t-

On the other hand Wagner's theory has had numerous supporters, for the weak point in
Midler's hypothesis is met by it, as many insects with very few corneal facets have
evidently extremely acute vision. The absence of any retina in the position of the
image below the cornea is, however, fatal to it.

Another difficulty militates strongly against this theory, which has been already dwelt
upon by Dr. Grenadier : the extreme minuteness of the subcorneal image would necessitate
recipient retinal elements far smaller than any known to exist in the animal kingdom.
Moreover, as each corneal image corresponds, as a rule, to an angle of about fifteen
degrees, and as the mean difference in the axes of adjacent facets is not usually half a
degree, the images of adjacent facets are nearly identical $, a fact in itself almost fatal to
Wagner's hypothesis.

Formerly I accepted a modification of Miiller's hypothesis § ; but a further investiga-
tion has led me to discard both the theories of arthropod vision, and to substitute one
which is, I believe, in complete accord with all the facts, and which, so far as I know, has
not been even suggested by any previous observer.

The compound eye is divided into two parts by a membrane which I shall speak of as
the membrana basilaris. The structures between this membrane and the cornea are the
crystalline cones and the great rods || ; I regard all these as dioptric structures, and shall
speak of the cornea, and all the parts which intervene between it and the membrana
basilaris, as the Dioptron. I shall use the term great rod provisionally to designate the
complex structure between each cone and the membrana basilaris, and segment of the
dioptron for the parts, collectively, beneath each corneal lens.

Beneath the membrana basilaris I have discovered a layer of bacilla, comparable in
most respects with the rod and cone layer of the vertebrate, in the place where Gottsche
suggested such a structure might exist. This layer is succeeded by others, similar to
the nuclear and molecular layers of the vertebrate eye. I shall speak of all these struc-
tures as the Neuron.

The existence of a continuous basilar membrane, in some arthropods at least, excludes
the great rods from any share in the nervous mechanism, a conclusion which also
follows from a more complete knowledge of their structure.

Nearly two years ago I made an examination of the eye of a Plume Moth (Ptero-
flwrus pentadactylus), in normal saline solution, immediately after the death of the
insect. I was surprised to find a very considerable deviation from any structure previ-
ously described by others or observed by myself. Fig. 3 is a representation of the
optical section of a portion of the eye : in this condition the great rods presented the

* Miill. Arch. 1852, p. 434.
§ Phil. Trans, ibid. p. 577.

f L.c. % Phil. Trans. 1878, pt. ii. pp. 596, 597.

I| " Rhabdia" mihi, Phil. Trans. I.e.

58*

392 ME. B. THOMPSON LOWNE ON THE COMPOUND VISION

appearance of ovoid spindles, each enclosed in a cylindrical sheath. The spindle-shaped
bodies underwent rapid change of form from the escape of their contents, so that in a
short time they were reduced to the condition of empty shrivelled tubes.

Further observations on various insects have convinced me that such changes usually
occur shortly after death, either from osmosis or from alterations of tension.

These observations suggested to me the possibility that the spindles of the great rods
should be regarded as lenses of very short focal length, but of great thickness, and that
they form a second refractive system, the anterior foci of which correspond with the
subcorneal images, and the posterior foci with the bacillar layer of the retina.

According to this view the dioptron is composed of an anterior and a posterior refrac-
tive system.

The anterior refractive system of each segment of the dioptron consists of a corneal
facet, a lenticulus, to be hereafter described, and the anterior face of the crystalline cone.

These parts form a subcorneal image, which lies in the anterior focus of the posterior
refractive system.

The posterior refractive system magnifies the subcorneal image, and erects it, at its
posterior focal plane, upon the bacilla of the neuron.

It is well known that if an objective is placed in the reversed position beneath the
stage of a microscope, and the instrument is focussed for its posterior focal plane, it can
be used as a telescope. A segment of the dioptron of an arthropod's eye, according to
my view, is comparable to such an instrument, and the whole dioptron to as many
instruments as there are segments, each giving a perfect picture of the objects which
subtend a small angle with its axis ; and thus a mosaic of images, which are erect and not
reversed, falls upon the retina.

Although the ordinary simple eyes of insects do not exhibit any structure comparable
to the second refractive system of the dioptron, the simple eyes of many larval insects
have a posterior refractive system, and afford a complete transition from a simple to a
compound eye.

I shall now proceed to the consideration of the details of structure and measurement
on which my theory is based, in the following order : —

I. The Anatomy and Functions of the Dioptron.
II. The Anatomy and Functions of the Neuron.

III. The Development of the Compound Eye.

IV. The Morphology of the Eyes of Arthropods.

I. The Anatomy and Functions of the Dioptron.

The dioptron in the most highly specialized forms of compound eye, such as are found
in the majority of perfect insects, is entirely enclosed in a chitinous case, formed by the
compound cornea and the basilar membrane, united to each other by an inflection of
the integument, which forms a short hollow cone between them.

The cornea and the basilar membrane are nearly parallel surfaces, so that the whole

AND THE MOEPHOLOGY OF THE EYE IN INSECTS. 393

lioptron has the form of a round or oval truncated cone, of which the base and apex are
iuhspheroidal or, more generally, cycloidal surfaces*. The older writers named the
nflected ring which unites the cornea and the basilar membrane "t7ie sclerotic" ; I shall
lesignate it the scleral ring.

The inner margin of the scleral ring is often thickened, and receives the insertion of
me or more muscles (figs. 1 & la. mc), which are probably concerned in the adjustment
accommodation ?) of the optical apparatus.

The whole interior of the dioptron is divided more or less perfectly into hexagonal
>r square tubes by fine sheaths f which extend from the margins of the corneal
acets to the basilar membrane ; each encloses a cone and a great rod, and forms a
segment of the dioptron.

The great eyes of the dragon flies (JEshtiia and Agrion) have enabled me to detect
arge lymph-sinuses (a.b. figs. 5 & 9) which jtass through the scleral ring and connect the
ymph-spaces of the dioptron with those of the head.

In JEshnia the afferent vessel (a) enters the superior internal border of the dioptron : it
las distinct walls, and is the anterior extremity of the bifurcated aorta %. I am unable
:o say that this is general in insects, but from what I have seen I suspect that it is.

The efferent openings (b.b. figs. 5 & 6) are situated in the lower portion of the scleral
ring, nearly opposite to the afferent vessel : they are generally, even in very different
insects, long slits, across which a number of radiating nucleated fibres rf are stretched,
between the basilar membrane and the scleral ring. These in sections, vertical to the
basilar membrane, present a fan-like arrangement similar to that of the ciliary muscle
of the vertebrate.

I have been unable to make out any stride in these fibres ; and as non-striated muscles
are not known to occur in insects, I hesitate to regard them as contractile, although one is
almost tempted to such a belief from their arrangement and position ; perhaps they are
simply elastic bands which maintain the tension of basilar membrane, where its attach-
ment to the scleral ring is defective, to permit the circulating fluid to pass out of the
dioptron. Their occurrence, however, in the simple eye (fig. 33, Is.) around the edge of
the corresponding membrane is indicative of an active function.

The tracheal vessels (t.v.) of the dioptron consist of a main trunk, which almost surrounds
the edge of the basilar membrane, from which numerous branches ramify and anasto-
mose on the neural surface of the membrane (fig. 68).

Vessels from this network perforate the basilar membrane and run outwards, almost to
the cornea, between the cuticular sheaths of the segments of the dioptron ; they termi-
nate in blind extremities.

In the SyrphidcB the tracheae of the dioptron assume the form of fusiform sacs
between its segments ; these have very narrow necks where they pass through the basilar
membrane.

The basilar membrane is a cuticular structure, which was described by Leydig as a

* Phil. Trans. I.e. p. 595. t " Umhiilluiigsschlauche " of Leydig & authors.

t ' Anatomy of the Blowfly,' by the author, plate is. fig. 1.

394. ME. B. THOMPSON LOWNE ON THE COMPOUND VISION

fenestrated membrane *, and it is usually believed to be perforated for the passage of
nerve-fibres to the great rods. In many sections it is easy to see that the membrane is
continuous, except where it is perforated by the tracheal vessels. In the great Dragon-
flies it is also undoubtedly pierced by the fringes from the pigment-cells of the dioptron,
which intercommunicate with the fringes of the pigment-cells, situated on the neural
surface of the membrane (fig. 69). In some insects the membrane is actually thickened at
the inner terminations of the great rods, so as to form small lenticular swellings (fig. 71);
and in a specimen of the eye of a lobster, in which the neural elements beneath the great
rods are not in the same line with them, the swellings are prismatic (fig. 70).

In some insects, especially in Notonecta, the basilar membrane is strongly ridged, the
ridges corresponding with the attachment of the cuticular sheaths of the segments of the
dioptron. It is apparently defective in parts, the openings being closed by the cellular
layers which cover the two surfaces of the cuticular membrane.

In the crane flies f (Tipula), in some Coleoptera (Telephorns), and in other insects
in which the component segments of the dioptron resemble distinct ocelli rather
than parts of a compound eye, the cuticular sheaths of the great rods extend inwards,
and also include the nervous structures. The basilar membrane then appears to be
actually perforated, so that the inner extremities of the great rods come into contact with
the nerve-terminals. This condition, however, is very exceptionnl, and is capable of
explanation from the manner in which the parts are developed.

I shall recur to the consideration of the question of the passage of nerve-fibres through
the membrane when I discuss the function of the great rods and the bacillaof the neuron |.

The various modifications of the cornea will be more conveniently considered here-
after ; I shall only mention in this place the fact that in all those cases in which there
are no lenticular facets, and where the cornea is simple and continuous, the outer face of
the crystalline cone is strongly curved, being a portion of a prolate spheroid, and there-
fore probably capable of producing an image of great sharpness. Such a surface has, as
is well known, the property of forming a perfect image, when the eccentricity of the
generating ellipse is the reciprocal of the refractive index, a condition which appears to
me to be very nearly attained in the curvature of these cones.

Prolonged investigation has gradually convinced me that the very considerable diffe-
rences which the plan of the arthropod eye exhibits in different species and families
are due to differences in the consistency and chemical nature of the parts. The highly
refractive structures may consist of some modification of chitin, or some allied albumi-
noid, or of an oil-like fluid contained in the meshes of a fine stroma and enclosed in
elastic capsules.

In the former case, where the refractive medium is chitinous, it undergoes but little
change of form in the preparation of microscopic sections : in the latter case the fluid
escapes from its capsule, and the whole appearance of the parts is modified, even when
such solvents as clove oil and absolute alcohol have not been used ; and the appearances
are still further altered by the use of such fluids in the preparation of sections.

* Miillcr's Arch. I.e. t Phil. Trans. 1873, l.c. p. 570. t See page 409.

AND THE MORPHOLOGY OF THE EYE IX INSECTS. 395

In most insects a lens lies under each corneal facet ; I shall show hereafter that this
3ns is sometimes formed from the outer portion of the crystalline cone, and in other
ases from the cornea itself, which then assumes the appearance of a honeycomb, the
ells containing the lenses.
The consistence of the lens seems to vary very much ; in some cases it is apparently
uid, enclosed in an elastic capsule ; in the Earwig and the Cockroach I have even been
ble to separate such lenses from the cornea by pencilling its inner surface ; the capsules
an then be ruptured, and the fluid seen to escape by pressing on the cover glass ; the
uptnred capsules exhibit a single tear and fine wrinkles (fig. 21) *.

The refractive index of the fluid with which the lenses are filled is very nearly 2 0 ;
his fluid appears like an oil, but it undergoes slow solution and decomposition by the
ction of water and weak saline fluids ; a comparatively lowly refracting fluid and a fine
eddish molecular substance result from their action.

The oil-like fluid is rapidly dissolved by ether and is blackened by osmic acid. Only
he red molecular precipitate remains in specimens prepared in the usual way with clove
til. I believe the great brilliancy of the cornea of many insects during life is due to
his fluid lens immediately beneath its surface, and that the loss of brilliancy which
tccurs soon after death, is due to the decomposition of the fluid, or its escape from the
ens-capsule.

I am inclined to regard this fluid as an oil of complex constitution, which is possibly
■ich in sulphur or phosphorus, to which it owes its high refractive index and ready
lecomposition. Observations with a micro-spectroscope have given negative results.

In other cases the lens, when separated, breaks up like a viscid body ; the lens-capsule
dways appears to contain a stroma, the meshes of which the fluid permeates ; and the
consistence of the lens depends on the relative amount of the stroma.

Every one who has examined the compound eye since improved methods have been
idopted, must have been puzzled by the " nuclei of Semper " which figure so promi-
nently in the descriptions of Claparede f. I formerly supposed with Dr. Grenadier that
Claparede so designated the nuclei which are frequently seen in the immature eye
between the crystalline cone and the facet of the cornea (en. tig. 3). But such a view is
by no means satisfactory.

I was very much surprised on one occasion, to see the " nuclei of Semper," which are
really nothing but the shrivelled segments of the lens-stroma, appear suddenly in a
compound cornea treated with ether on the stage of the microscope ; as the oil dis-
solved out of the lenses the contents of the capsules split into four parts. I regard
this as the result of the shrinking of the stroma. A similar appearance is seen when the
cornea of an insect is examined after having been allowed to become partially dried. I
have never observed this in specimens prepared in the usual manner with alcohol and
clove oil.

In the simple eyes of caterpillars , " ocelli coitijjosili " (figs. 36-10), which are un-

* This lens was described by Miiller ; but its existence has been completely overlooked bv reeont authors,
t Zeitschr. f. w. Zool. Bd. x.

396 ME. B. THOMPSON LOWNE ON THE COMPOUND VISION

doubtedly nearly related to the compound eye, a subcorneal lens exists. In these it
consists of three segments. A similar lens is seen in the semicompound eyes of isopods,
and is regarded by Dr. Grenacher as a cone in a highly modified condition. It
consists of only two segments instead of three or four.

The cornea and oil-lens together have a very short focus. The picture formed by
these structures is usually from five to ten micromillimetres behind the posterior surface
of the lens.

Very great differences of opinion have obtained credence concerning this image since
the publication of Gottsche's paper. Many authors, and quite recently Exner *, have
denied that any images are formed by the eye when the crystalline cones are in situ. I
have a specimen of the cornea and cones of a moth (Smerinthus populh), mounted in
balsam, in which distinct images are formed in the interior of the crystalline cones. In
Exner's experiment the crystalline cones were surrounded by a lowly refracting fluid.
As the images, according to my calculations, are formed near the focal plane of the
inner ends of the crystalline cones, rays emerge from the inner extremities of the cones
as parallel rays, or at least approximately parallel. The image cannot therefore be
observed by the high powers of a microscope. Exner is undoubtedly quite right when
he says that no image can be observed when the cones are in situ and surrounded by a
lowly refractive fluid ; but it by no means follows that none is formed. Indeed the
structure of the eye is such that an image must be formed, and it cannot be neglected in
working out the manner in which the eye acts as an organ of vision.

The focal length of the corneal lenses of the Fly (Jfusca vomitoria) has been given by
me from actual measurement as ^q of an inch f . I have since found that the results
obtained by measurement are always greater than those arrived at by calculation ; this is
partly due to the rapid loss of refractive power immediately after death, but also to the
fact that any moisture adhering to the inner aspect of the lenses forms a concave surface
as it is attracted by the inflected margins of the corneal facets. I believe that the only
satisfactory results are those arrived at by calculation. The greatest difficulty is the
estimation of the refractive index. This is often considerably greater than that of
Canada balsam, which is sufficiently proved by the formation of a subcorneal image
when the cornea and cones of a moth are entirely immersed in it.

Small fragments of glass in water have nearly the same brightness as the cones and
lenses of a moth in balsam. By reducing the illumination of the field of the microscope
both were just visible with the same illumination, which indicates a refractive index, for
the cones and lenses, of 1*8 nearly. Assuming the refractive index to be between 1-5

and 2-0, the focus by the formula ( ^3)7x7=/ will lie between r and '2, when r—r'.

the value of the fraction ^-p varies between 10* and 20* in different insects which I have

examined. Therefore the focal length of the corneal lens is from five to twenty micro-
millimetres, and falls well within the crystalline cone, and the rays frequently emerge

* Biol. Centralblati, i. p. 280.
t Phil. TraDs. I.e. p. 585.

AND THE MOEPHOLOG-Y OP THE EYE IN INSECTS. 397

from the posterior end of the cone as approximately parallel, a conclusion which I
arrived at formerly *.

In a specimen of the cornea and cones of the moth {Smerinthus populi) mounted in
balsam, the micrometer screw shows the distauce of the image behind the posterior
surface of the corneal lens to be about forty micromillimetres, estimating the refractive

index of the lens at l-68 and that of the balsam at 1/5 .\ -r-r- = ri2, and 40 : t-ttt '■ '■ f :

1"5 1'12 J

■pgg .•./=26 micromillimetres approximately. The radius of curvature of the anterior

surface of the corneal lens is about forty micromillimetres, and the distance between its
anterior and posterior surfaces is 10 micromillimetres, which would give a focal distance
of 26 micromillimetres measured from the posterior surface as before, with the same
refractive index.

The diameter of the image is approximately 20 micromillimetres, in the same balsam-
mounted specimen ; it corresponds to an angular aperture of 24 degrees. The distance
is 10 X cot 12°=47 in micromillimetres behind the optic centre ; which does not differ
widely from the above. The angle was determined by using two lighted tapers as the
object. The error arises from the difficulty of determining the exact size of the image,
that is the distance between the focal points. Of course the plane mirror of the micro-
scope was used in making these measurements.

I formerly t gave the approximate length of the focus of the corneal facets of the
hornet as -»\q of an inch ; it is easy to see that it should have been -^oo, with a
refractive index which I estimated at 1'53, which is too low, as the refractive index is
nearly 2%0 during life — a result which is sufficiently near those given above.

If we assume the outer ends of the great rods to have a spheroid curvature convex
towards the cone, the posterior focus of which corresponds to the bacillar layer, all the
mechanism is present for the formation of an erect magnified image of the central
portion of the subcorneal inverted image upon the sentient structures.

In many insects I have observed such a conformation of the outer ends of the great
rods as this theory requires, and I believe that when observation fails to show such an
arrangement, this is due to the very profound modifications which these structures
undergo when removed from the eye, or during the preparation of the eye for investi-
gation.

J. Muller regarded the great rods as nerve-terminals, a conclusion which was justified
by his want of knowledge of their structure, but which is no longer tenable. Gottsche
first discovered the compound nature of the great rods, and described, as I have already
mentioned, the inner extremities of those of the lobster, as " double pyramids." The i
structure to which he gave this name is now better known as the spindle; Dr. Grenadier/
speaks of it as a " Retmula."

Max Schultze examined the structure of the great rods and their spindles in the
lobster ; he came to the conclusion that the rods are not the receptive elements, but
ascribed tbis function to the spindles. Max Schultze founded this opinion chiefly on

* Phil. Trans. /. c. p. 581. t Phil. Trans. /. c. p. 581.

SECOND SERIES. — ZOOLOGY, VOL. II. 59

398 ME. B. THOMPSON LOWNE ON THE COMPOUND VISION

two characters which he had observed in these structures — their tendency to split into
transverse disks and their pink colour in the lobster.

The structure and development of the spindles varies considerably in different Arthro-
pods. Sometimes they occupy the whole interior of the great rods, extending from the
basilar membrane to the apex of the cone, as in the flies, and more obviously in the eye
of the Water Boatman (Notonecta) ; in a greater number of insects, however, they are
separated from the cones by a considerable space, the interval being occupied by an
albuminous fluid or semifluid substance enclosed in a cuticular sheath, which is sur-
rounded by pigment-cells (fig. 22. ch.).

The albuminous contents of this cuticular sheath have a low refractive index, and
partially enclose both the cone and the spindle. The diameter of the tube which inter-
venes between the cone and spindle is subject to considerable variation in the same species
of insect. In moths it is usually a very fine thread in specimens prepared by section ; but
in the recent eye of the very species in which it appears in this form in sections, I have
frequently found the tube as wide as the base of the cone (figs. 3 & 27). I believe that
the contracted condition is produced by the escape of its contents and the great elasticity
of the cuticular sheath itself.

In the perfectly fresh eye of an insect the spindles are very transparent ovoid bodies,
attached by their bases, which are truncated, to the membrana basilaris, and surrounded
by pigmented fringes from cells, which cover the outer surface of the basilar membrane
between the spindles (figs. 2, 3, & 10. sp.).

The appearances which the spindles assume after death and in sections prepared for
microscopic examination are very various. In the noctuid moths they then appear as
chitinous rods ; sometimes they even present a stellate transverse section. In a specimen
of the eye of a moth (Hemerophila perfumciria), prepared with osmic acid and afterwards
mounted in balsam, they have assumed the form of absolutely empty shells, which are
blackened intensely by the acid.

The spiudles of a yellow Underwing (Triphcena pronuba, fig 22 sp.) from a specimen
preserved in a solution of chloral hydrate which was not replaced by any other fluid,
resemble tubes filled with minute spherical granules, which give them a transversely
striated appearance. These spindles are white and opaque when seen with reflected
light.

In all these cases the form and appearance of the spindle are very different from that
which it presents in its normal unaltered condition. The outer end always contracts
much more than the inner, so that it has a strong tendency to assume the form of a wine-
bottle with a long neck. In the Crane-flies (Tipnla) I have found that the length of the
narrow part of the spindle varies greatly in different specimens (figs. 10-13), so that there
is no difficulty in connecting the normal ovoid spindle with the bottle-shaped organ
which is usually seen in microscopic preparations.

The spindles of the Crane-flies (Tipula) are seen, in transverse sections, to be composed
of seven tubes (fig. 11).

In some recent preparations I have succeeded in isolating these (fig. 13) ; they then
curl and twist in a very remarkable manner, in water and glycerine, and they are easily

AND THE MOEPHOLOGY OF THE EYE IX l\si;< 399

ruptured by pressure ou the cover-glass, when a highly refractive oil-like fluid escapes
from them. The axial structure of the great rods in other insects [Notonecta) exhibits
the same appearances, some of which are figured by Dr. Grenacher*, and I have no
doubt that the very fine axial threads seen in the great rods of dies arc the shrivelled
remains of the spindle.

In the true flies (Musckke) the great rods are very highly modified ; each consists of
six cells (sheathing-cells) ; these enclose the axial spindle. This can only be satisfactorily
examined, in the recent eye, by very careful teasing. The great rods may be thus sepa- \
rated either in normal saline solution, or in very dilute osmic-acid solution, -05 per cent. ;
in the former fluid the sheathing-cells undergo very rapid disintegration by vacuolation,
but this occurs less rapidly in the osmic solution.

The sheathing-cells surround a structure which is fluted like a column, the " Rhab-
dom" of Dr. Grenacher. This is the spindle; it consists of an elastic sheath, apparently
composed of from four to seven component tubes filled with a highly refractive fluid.
The fluid rapidly escapes from the injured tubes in minute glistening drops, whilst tin
tubes themselves become converted into fine threads. In the recent condition the
slightest pressure suffices to empty the spindles. The various appearances of the com-
ponent tubes are represented in the figures (figs. 16-18).

The tubes of which the spindle is formed apparently intercommunicate at the inner
portion of the spindle. They twist and curl in glycerine and water as well as in normal
saline solution. They are very elastic, and are easily drawn out to several times their
proper length, which they regain when the tension is relaxed.

The fluid contents of the spindle are soluble in ether and clove-oil. The fluid is
perfectly colourless in the Blow-fly; but the same fluid in the cabbage butterfly has a
bright ruby tint when seen through the length of the spindle.

In ordinary transverse sections (fig. 20) the spindle is seen to contain a cavity sur-
rounded by a plicated wall. The boundary of the irregular lumen exhibits six or seven,
more rarely four, bright spots : these are the sections of the threads. All these details are
correctly figured by Dr. Grenacher f. The outer extremity of the spindle in the Hies [s
nearly hemispherical, and is imbedded in a structure which I formerly described as a
C tetrasome. This structure, in preparations hardened in chromic acid and its salts, has the
appearance of four small spheroid bodies ; but in some remarkably distinct sections which
were fixed with osmic acid these are seen to be the four segment sofa minute lens. The form
and relation of this structure to the spindle will be best understood by a reference to fig. 19.

In the Lepidoptera the distance of the external end of the spindle from the cone
appears to depend on the habit of the insect. It is furthest removed in the Noctuids,
and closest in the Diurna: that is, in the Noctuids it is a larger lens of longer focal
length, and therefore it receives a much larger cone of light-rays; in the Diurna the
body which I formerly named the "tetraphore" $ appears to be the external portion of
the posterior refractive system.

In the Hymenoptera and Neuroptera, it is very difficult to sec the parts in their normal

* L. c. figs. 51-53. + L. c. fig. 62. t Thil. Trans. /. 1 .

ry.i

400 ME. B. THOMPSON LOWNE ON THE COMPOUND VISION

condition, as the walls of the refractive structures are very thin, so that in all the speci-
mens which I have examined, the tubes of the spindle, which is very similar to that in
the flies, had collapsed.

There are usually two principal layers of pigment in the dioptron — an outer layer,
which is thickest at the apices of the cones (fig. 1. fgx), and an inner layer (pg3) at the
inner extremities of the spindles.

The pigment of the outer layer is generally contained in cells which form a kind of
iris around the inner extremity of each cone. It will he convenient to designate these
rings of pigment-cells "irides," in conformity with the nomenclature of the older writers.
There are, as a rule, five or six iris-cells around each cone. These send pigmented
fringes outward over the cone and inward over the great rods. In many insects
(Hymenoptera and Nematocerous Diptera) the outer set of fringes are replaced by rod-
like prolongations of the iris-cells. Each cell has a single pigmented rod, which has a
bright, highly refractive spherule at its outer end, immediately beneath the cornea.
These may be called ciliary rods (figs. 31 & 32. cb).

I have found ciliary rods in all the Hymenoptera and Nematocerous Diptera which I
have examined.

In the Dragon-flies the soft cone is surrounded by pigmented fringes. Amongst these
there are from twelve to sixteen thread-like processes which are not pigmented ; each
terminates in a bright, highly refractive spherule, like that of a ciliary rod. These
spherules form a ring around the cone immediately beneath the corneal facet. They are
figured by Claparede *. I shall speak of these threads as " ciliary threads."

The form and connexion of the ciliary rods and threads suggest the idea that they
may be concerned in regulating the amount of light which passes through the apex of
the cone by means of a local mechanism. The action of light on the highly refractive
particles possibly gives rise to a contraction of the iris. Whatever the mechanism by
which the size of the aperture in the iris is regulated, it is certain that it is contracted
in a bright light and dilated in the dusk, in nocturnal insects at least.

In Moths the well-known luminous reflex of the eye after the insect has been kept
for a time in darkness is undoubtedly due to reflection from the spindles, as these are
surrounded, in them, by numerous very close parallel tracheal vessels, which form a
very perfect reflector. The disappearance of the reflex in the light is certainly due to
the contraction of the irides. The gradual contraction of the bright field in each lenti-
culus of the cornea can be observed by means of a modification of the ophthalmoscope
which I devised and shall hereafter describe.

It is a remarkable fact that the irides of the peripheral segments of the dioptron close
before those of the central portion when the light falls directly in the axis of the latter,
so that when the luminous eye of a moth is observed, it appears as if the bright part
gradually became smaller until at last it entirely disappears.

Kiihne is certainly mistaken when he says that the luminous reflex is of periodic
recurrence f , and cannot be produced in the day. After exposure to a bright light the

L. c.
t W. Kiihne, " Observations on Notodon and Acherontia," Untersuch. Phys. Heidelberg, Bd. i. p. 242.

AND THE MOEPHOLOGY OF THE EYE IX INSECTS. L01

iricles are often a very long time before they dilate, but I have never failed to sec the
reflex after the animal has been kept in the dark for half an hour, although it disappears
almost instantly even in diffused daylight.

In the flies (Muscidse) the cone is surrounded by four large flat cells, which extend
from the margin of the cornea to the iris. All trace of ciliary rods and threads is
wanting, although the irides are well developed.

The inner layer of pigment-cells is in immediate contact with the basilar membrane;
the cells form rosettes around the inner extremities of the spindles, and send pigmented
thread-like processes over their surface. In the Dragon-flies these cells give oil' a number
of pigmented fringes which perforate the basilar membrane, and communicate with
similar fringes from the pigment-cells of the neuron.

In the flies and in some Hymenoptera an intermediate set of stellate pigment-cells
exists {pg-) between the inner and outer pigment-cells of the dioptron. These cells inter-
communicate with each other, and in some preparations this gives the great rods the
appearance of branching and uniting with each other. I think that this appearance is
usually deceptive: the spindles are certainly always distinct, hut appearances in sections
of the eye of the Lobster, and in some osmic-acid preparations of the eye of a Moth
(Hemerophila perfumaria), certainly indicate that the sheaths of the spindles do inter-
communicate in some parts of the eye.

I shall conclude my description of the dioptron with an account of some of the modi-
fications in its several parts which I have observed in different insects, and I shall add
what I am able concerning their development.

I feel that in many particulars this account will he somewhat incomplete, especially
with reference to the development of the several structures ; but I am unwilling to
delay this communication for a fuller investigation, which will certainly occupy several
summers at least.

A. Modifications of (he Cuntea mid Lens.
The compound cornea exhibits four distinct types in insects, three of which arc seen
in the various stages of development in the eye of the Common Cockroach. All the
are also found in the perfect condition in other Arthropods. In the earliest stage of
development in the larval Cockroach the cornea is simple and continuous, without a
' trace of facetted structure. In this condition it consists of several transparent cuticular
layers (figs. 75 and 76. c). The entire cuticular cornea is shed with each ecdysis, and is
renewed from a layer of cells which, prior to the ecdysis, is situated between the old
cornea and the cone. I believe that the new cuticular cornea is formed heneath, and o
on, the surface of this cellular layer. In a larva of a Dragon-fly {Agrion), half an inch
long, I found a layer of large flat hexagonal cells covering the whole surface of the
cuticular cornea, one cell corresponding to each segment of the eye. These cells were
firmly united to each other by their edges, and each contained a large lenticular nucleus.
In a section of this cornea the cuticular layers are seen to he very numerous, and they
lie beneath the cellular layer. The nuclei of the cells stain very readily with eosin,
whilst the remainder of the cornea remains unstained. The outer cellular layer is usually

402 MR. B. THOMPSON LOWNE ON THE COMPOUND VISION

absent when the cornea is continuous and without facets. I suspect it is shed after the
development of the cuticular layers beneath.

I have named this type of cornea the simple non-facetted cornea. It is apparently
the permanent condition in Gammarus and Branchippus amongst the Crustacea.

The second type of cornea (fig. 30) is seen in the functionally active eye of the larval
Cockroach. It is divided into lenticular facets. I shall speak of this type as the
facetted cornea.

The facetted cornea exhibits two distinct layers — an outer nuclear layer and an inner
laminated layer.

This form differs from the continuous cornea in the persistence and further develop-
ment of the cellular layer. The large flat nuclei of the cells enlarge at the expense of
the cell-substance until they form the whole surface of the cornea, one nucleus forming
each facet. This layer, when fully developed, is very hard and brittle. It is no longer
stained by eosin, as in its semideveloped condition, but assumes a yellow colour in
specimens which have been treated with chromic acid or its salts.

The structure of the facetted cornea is worthy of study in dried insects. I have a
beautiful section from the eye of an African Carpenter Bee (fig. 30). The lenticular
nuclei are firmly cemented together, and exhibit no trace of minute structure. The
laminated portion beneath these consists of from fifty to sixty layers. These follow the
contour of the inner surface of the lenticular facets.

Fine lines are also seen running perpendicularly to the lamina?. These correspond to
the divisions between the facets. There are also distinct lacunse between the laminse,
with minute teeth on each side of the lacuna ; indicating that the laminated structure is
probably due to the existence of finely serrated fibres interwoven with each other.

The third modification of the cornea in the Cockroach is only seen in the final stage
of development : I shall speak of it as the histoid cornea * (figs. 3 & 10).

I have not met with any description of the kistoid cornea, although it is preeminently
characteristic of the most highly differentiated forms of compound eye. It consists of a
chitinous (?) cuticular membrane folded on itself so as to resemble a piece of honeycomb,
the openings of the hexagonal cells being turned inwards. The closed ends of the cells
are segments of spheres : these form the surface of the cornea.

The hexagonal walls of the cells are often highly chitinized and deeply pigmented ;
the curved ends which form the corneal facets are thin, membranous, and very trans-
parent. The material of which they are formed is usually soft, flexible, and elastic, and
the two surfaces of the membrane are parallel to each other.

The cavity of the corneal cell is occupied by the oil-like lens which I have already
described.

This lens is certainly developed from the substance of the cornea. I was fortunate
enough to obtain sections of the eye of a Blow-fly which throw light upon this point.
Fig. 15 shows three segments of the dioptron of an immature imago, which had not long
before escaped from the pupa.

* KtVrr,, a capsule or small case.

AND THE MORPHOLOGY OF THE EYE EN INSECTS. 10:5

Each corneal facet consists of a chitinous membrane overlying a soft substance which
is easily stained with logwood or carmine solution.

As development progresses the deeper portion of the cornea becomes more highly
refractive, and loses its property of taking up status. It appears as if the albuminous
substance of the young cornea becomes gradually converted iuto a stroma infiltrated
with oil. I have never seen the oil in drops, like those in a young fat-cell. In the
rudimentary lens it appears rather as if the oil were evenly distributed throughout the
whole substance of the lens.

In the condition which I have just described I have seen no traces either of division
of the albuminous portion of the cornea into four, or of four nuclei; but in a still
earlier stage there are distinct indications of four nuclei in each corneal facet, and in a
later stage the appearance of the "nuclei of Semper" on the addition of ether, or by
drying, points to the origin of the corneal lens from four original cells.

In the immature insect, during the development of the eye, a number of capsules,
filled with cells containing an abundance of oil in large drops, are seen under the mem-
brana basilaris. Similar cells are shown in some of Claparede's figures, and arc named
by him basal cells * . These are entirely absorbed during the development of the eye,
and it appears probable that they afford the oil required for the formation of the refractive
media. I have no doubt they belong to the system of fat bodies in which the larvae and
pupse of insects are so rich.

The fourth form of cornea is apparently confined, amongst insects, to the imago form
of some Gnats, but is also probably the condition in Mysis, in its perfect state, and in
some other Crustacea. I propose to distinguish this type of cornea as the lenticular
cornea.

The lenticular cornea consists of a number of well-developed, almost hemispherical
lenses, which are nothing more than the united crystalline cones of the larva and nymph.

Dr. Grenacher has correctly figured and described the eye of the larva and pupa of a
Gnat; but be adds : " I have only studied the eye of this insect (Corethra plumicornis)
in the larva and pupa, not in the imago;" and says, "This is a matter of no importance,
as the eye undergoes no change in passing from the pupa to the imago " t. 1 was at
once struck by this remarkable statement, as the eye is covered by a smooth non-facetted
cornea in the nymph, whilst the corneal facets in the imago are remarkable for their size
and very great convexity (fig. 74)— a fact well known to all who have made the most
cursory examination of the compound eyes in this family of insects.

A comparison of a section of the eye of the adult larva (fig. 70) or nymph (fig. 75)
with that of the imago (fig. 74) shows that the great convex facets of the cornea of the
perfect insect are really the ovoid crystalline cones (" spherocones") of the larva, covered
by a thin cuticular membrane, which dips down between them, as the membrane of the
kistoid cornea dips between its lenses.

In some of the Gnats the outer nearly hemispherical segment of the cone becomes
converted into a nearly globular oil-lens, exactly bke the oil-lens of a Crane-fly. The

* L. c. t /- ■ p. 94.

I

404 ME. B. THOMPSON LOWNE ON THE COMPOUND VISION

inner segment of the cone then hears the same relation to the other refractive structures
that the cone hears in the ordinary Diptera and Hymenoptera, or in insects and Crus-
taceans generally.

The formation of a fluid lenticulus from the suhstance of the cornea is not confined to
the compound eye, hut is also seen in the very different simple ocelli of some Arachnids.
In a section of the simple ocellus of a Spider (Salticus scenicus) (fig. 34), which I pre-
pared in the usual way and mounted in Canada balsam, the corneal lens is absolutely
hollow. It evidently contained a fluid during life.

The sudden scintillations, which the bright reflex of the fundus of the eye of this insect
exhibits in the living state, seem to indicate a power of accommodation, possibly from
the action of muscle fibres which are inserted around the edge of the cornea.

Dr. Landois * traced the development of the subcorneal lens in the compound ocellus
of a Caterpillar to a very different source. Soon after an ecdysis he found three spheroidal
bodies very similar to those which I have described as a tetrasome in the eye of a larva
of Acridiwm f . These unite to form the lens.

From the manner in which the cornea is formed in Gnats, I think it probable that this
subcorneal lens becomes the cornea at the next ecdysis, and that a new lens is then
formed from subcorneal cells or nuclei. Although I have not observed either subcorneal
nuclei or cells in the eyes of caterpillars, both are very frequently seen in the eyes of
larval insects.

I suspect that the compound cornea is developed in a similar manner in some Ortho-
ptera, from the indications I observed in Acridmm, but I am uncertain on this point;
neither can I do more in the case of the Isopods than point out the very strong resem-
blance of the subcorneal lenses to those of the compound ocelli of caterpillars.

B. Modifications of the Crystalline Cone.

The outer portion of each segment of the dioptron, between the cornea and the external
layer of pigment, may be conveniently termed " the chamber " (fig. 10 ch.). It is formed
by the cuticular sheath of the segment, often thickened by a deposition of chitin ; and is
lined by fringes, pigment-cells, or ciliary rods.

Tlie chamber contains the crystalline cone (cc.), which in many insects is replaced by a
coagulable fluid, contained in four vesicles, and in others by four nucleated cells. Dr.
Grenadier % classifies the eyes of insects in three groups : those with a cone he calls
euconic, those with the albuminous fluid pseudoconic §, and those with four nucleated
cells aconic.

Up to a certain point my observations correspond with those of this distinguished
author. In all insects the contents of the chamber are developed from four primitive
cells || . This condition persists in the Crane-flies ( Tipula), even in the recently formed
imago (figs. 10 & 77) ; but I Lave failed to find these cells in the perfectly developed
insect, in which the chamber is filled with an albuminous fluid, and I believe that the
presence of the four nucleated cells is always evidence of immaturity. Dr. Grenadier's

* Zeitschr. fur w. Zool. Bd. xvi. f L. c. Phil. Trans. J L. e.

§ Hydroconic, rnihi, I. e. Phil. Trans. || This was observed by both Claparede and Weismann.

AND THE MORPHOLOGY OF THE EYE IN INSECTS. 405

pseudoconic type is, according to him, confined to the Diptera with short antennse. In
this, again, I cannot agree with him, as the Dragon-flies exhibit precisely the same con-
ditions, and in the imago of the Cockroach I find a similar semi-fluid cone. In all these
osmic acid coagulates the fluid, and even the eyes of flies exhibit a well-marked cone
when the preparation has been so preserved (fig. 28).

The elongated cones which have been observed in the eyes of many Arthropods,
extending from the cornea to the membrana basilaris (as, for example, in Hyperia *,
Typhis t , and some other genera]:), appeared to me for a long time to present great
difficulties with regard to the views which I now hold. Further investigation has con-
vinced me that these cones are artificially produced by the coagulation and drying of
the albuminous tissues of the cone and spindle, with their cellular sheaths. The evidence
of this is seen in numerous specimens of the eye of the Cockroach. In some of these
there are distinct elongated cones, whilst in others more successfully prepared the spindle
is well seen, and has the same form and structure as it exhibits in Notonecta and Tipula.

The crystalline cones of the Nocturnal Lepidoptera and of the higher Crustacea are
probably morphologically distinct from the structures already described. The highly
refractive cone in Nocturnal Lepidoptera is surrounded by a softer, or even fluid,
sheathing cone, which extends from the apex of the crystalline cone to the spindle (figs. 3,
26, & 27, sh). I regard it (the sheathing cone) as the representative of the cone in the
Diptera and Dragonflies.

Both the hard cone and its soft investing substance are divided into four longitudinal
segments, indicating their origin from the four cells of the primitive cone ; but I cannot
at present decide whether the hard cone is formed from the inner portions of these cells
or between them, as Claparede § thought not improbable.

In both Noctuid and Crepuscularian Moths, the optical transverse section of the cone
often appears to contain four groups of deep purple granules ; these, in optical longi-
tudinal sections, are seen to lie on the surface of the cone, and are probably deposited by
the decomposition of the fluid which surrounds the cone. In dried as well as in balsam-
mounted specimens the sheath is so closely applied to the cone that it appears to form
its outer portion, and the coloured granules therefore look as if they were imbedded
in the substance of the cone itself || . Unicellular organisms are not infrequently found
in the fluid contents of the sheath of the cone : these are undoubtedly parasitic

(ng. 27, z).

It is by no means easy to understand the contents of the chamber in the Diurnal
Lepidoptera in such genera as Colias and Vanessa, although in Pieris brassicce no diffi-
culty exists, the cone being similar to that of a Noctuid Moth. I have already figured
the remarkable modifications seen in the two former genera^]". I suspect that the
cells (c. c.) in my figure are the representatives of the cone ; and that the tetraphore is
the outer extremity of the spindle, which retains its ovoid form after the escape of the
fluid contents of the tube, and that the tube forms a narrow stalk supporting the chitinized

* Clapaxede, I. c. t id. $ Leydig. § L, c.

|| Perhaps the coloured beads which cover the cone in Notonecta have a similar origin.
If Phil. Trans. 1. c. fig. 35.
SECOND SERIES. — ZOOLOGY, VOL. II. 60

406 ME. B. THOMPSON LOWNE ON THE COMPOUND VISION

outer portion of the organ, which does not collapse like the rest of the spindle. This
view is supported by a comparison of the outer end of the spindle in the Fly (fig. 19)
with the structure in question.

C. The Great Bods.

The great rods consist essentially of the spindles and their cellular sheaths. The more
important modifications of these structures have already been described, in their relation
to the alterations which they undergo after death, and I have nothing to add further
with regard to the very remarkable modifications which have been described by myself
and others, except that they result from post-mortem changes.

The most important point in relation to the theory of Arthropod vision is, however,
the direction of the axes of the great rods. It is well known that these are often seen
to be strongly curved, even in the most carefully prepared sections, and this fact has
been brought into prominence by the opponents of Midler's hypothesis * . Such a curved
condition of the rods and spindles would be still more fatal to my own view. I have
therefore carefully investigated this point, and conclude that this curvature is the result
of changes of tension in the parts of the dioptron, and of the elasticity of the spindles.
Sections made through the entire eye, immediately after the death of an insect, show no
such curvature of the axes of the great rods when examined as opaque objects with low
powers ; yet the eyes of the same species exhibit very strongly curved rods when sections
of the eye are examined in balsam. The manner in which the isolated rods twist and
curl in all fluids shows that they must be powerfully affected by the action of the fluids
usually required for the preparation of sections, and their action could scarcely fail to
produce contortion of the great rods, even in the closely packed condition in which they
exist in the cavity of the dioptron. The curvature would chiefly affect those rods which
are nearest the periphery of the eye — a condition seen in all the sections which I have
examined. The slightest pressure on the cornea in the recent eye permanently distorts
the great rods. Hence it can hardly be expected that the axes of these structures would
be undisturbed, at least around the periphery of the eye, in specimens preserved in any
fluid which affects the normal tension of the parts.

I have made numerous attempts to determine the optical relations of the dioptric
structures of the compound eye by means of a modification of the ophthalmoscope, but
at present I have not been able to throw any further light upon the functions of the
great rods by this means ; except that the colour of the reflex obtained appears to
depend on the colour of the fluid contents of the spindle.

I have found the best method of examining the reflex to be the substitution of a
reflecting ophthalmoscope for the eyepiece of a microscope. By this means a bright
luminous spot may be observed as a real image in the tube of the instrument. A
quarter objective must be used, and the mirror of the ophthalmoscope must be strongly
illuminated. The microscope is then focussed so that a real image of the corneal facets
is seen between the objective and the eye of the observer. By bringing the object-glass
gradually nearer to the insect's eye the reflex will come into view.

* Exner. Biolog. Centralblatt, i. p. 273.

AND THE MOKPHOLOGY OP THE EYE IN INSECTS. 407

The reflex appears as a disk having a fiery glow in Moths, and as a bright ruby spot in
the Cabbage Butterfly. Sometimes six spots, surrounding a central spot, are seen in the
eye of this insect : perhaps these are diffraction-images. A similar appearance is seen
when the eye of this insect is observed by the naked eye, except that the spots are black.
The central spot is always opposite the eye of the observer, whatever the position of the
eye of the insect.

The reflex seen with the micro-ophthalmoscope is green in Tipula and bright yellow
in the Diurnal Flies. Coloured diffraction-fringes are usually present around the central
bright spot in both these insects ; but the central image is sometimes surrounded by a
perfectly black ring.

The manner in which the luminous reflex scintillates is very suggestive of an altera-
tion in the focal plane of the dioptric structures under the control of the insect.

II. The Anatomy and Functions or the Neuron.

The neuron consists of a large nerve-papilla, or of a series of papillge, which arise
from the side of the procephalic ganglion, and form a retinal expansion on the inner or
neural surface of the membrana basilaris (figs. 41 to 07).

The neuron may be conveniently divided, for purposes of description, into three parts
— the retina (rt.). the optic nerve (no.), and the optic ganglion (g.1) (fig. 41).

The retina consists of a layer of bacilla, supported by a complex neuroglia (fig. 56).

The bacilla present an inner (i.) and an outer segment (o., fig. 53), like those of the rods
and cones of the vertebrate retina. Both the inner and outer segments are very easily
destroyed, and these structures are best examined by teasing out the recent retina in a
very dilute solution of osmic acid, '05 per cent.

The outer segments are cylindrical or conical, highly refractive, easily stained by
chromic acid and its salts, but unaffected by logwood. In the Dragon-flies the outer
segments often split into transverse disks.

The inner segments are protoplasmic, and are easily stained with logwood and carmine.
They are richly supplied with very fine tracheal vessels (fig. 53). The inner and outer
segments are of nearly equal length, and measure from 25 to 50 micromillimetres each.
The outer segments are usually about 2" in diameter, and the inner are somewhat thicker.

The bacilla are usually arranged in bundles, which are bound together by the neuroglia,
each bundle corresponding to a segment of the dioptron. In the Cockroach and the
Blow-fly I have occasionally found some of the bacilla with the outer segment double,
like the twin cones of fishes (figs. 65 & 66) ; and in some of the bacilla from the latter
insect I have occasionally observed a lenticulus (fig. 64) between the inner and the outer
segments. This is not stained by osmic acid, and closely resembles the lenticulus
described in some vertebrate rods. I have never observed anything like the coloured
globules of the cones of birds and reptiles.

The inner segment is usually finely granular, and is continued inwards as the axis-
cylinder of a nerve-fibre.

The most remarkable modification of the bacilla is the separation of the inner and

60*

408 ME. B. THOMPSON LOWNE ON THE COMPOUND VISION

outer segments iu many larval forms, in the imago of Tipula and in most Noctuid Moths.
In these cases a long fine axis-cylinder passes from the outer to the inner segment
(figs. 46, 55, 57, & 67).

In the Noctuids a tracheal network is interposed between the inner and outer segments
of the hacilla, permeated by the fibres which connect them with each other. The
tracheae form a kind of tapetum behind the outer segments.

A similar separation of the inner and outer segments of the bacilla occurs in the
simple eyes of an Arachnid (Phalcmgium), which has been figured by Dr. Grenadier *.

It will be observed that the outer segments of the bacilla, which correspond in cha-
racter with the outer segments of the vertebrate rods, are turned towards and not away
from the refractive media. In this they conform to the usual condition in the inverte-
brate eye. I regard this as of developmental rather than of functional importance.

The outer ends of the bacilla are separated from the basilar membrane by a layer of
fine branching endothelial cells. These are frequently pigmented, and send pigmented
fringes inwards, which closely embrace the outer segments of the bacilla. The pigment
^is not deposited between the bacilla and the membrane, but only around the bacilla. I
regard this layer as the analogue of the retinal pigment. It is always black. The
pigment is entirely wanting in the Diurnal Flies, but is very abundant in the Lepidoptera,
Hymeuoptera, and Neuroptera.

The neuroglia of the retina consists of fine fibres (fig. 56), connected with the basilar
membrane and with two distinct layers of small stellate cells — an outer layer between
the outer and inner segments of the bacilla (a), and an inner layer at the inner extre-
mities of the inner segments (»r). In many specimens a number of small granules are
seen crowded together between the outer segments of the bacilla (fig. 59). I am not
certain of their nature, but it has occurred to me that they are possibly the broken outer
ends of the bacilla themselves.

Sometimes large soft granular cells are seen in the same position (fig. 58). I think
these are an indication of immaturity.

None of my predecessors in this investigation have described the bacilla. It may,
therefore, be fairly asked, If these structures exist, how is it that they have been so
frequently overlooked ?

I think the answer to this question is found in the fact that it is only in very thin
sections, such as they had not the means of preparing, that the bacilla can be recognized.
In all ordinary sections a thick band of pigment between the membrana basilaris and the
optic nerve is all that can be made out. When I wrote my former paper I had only
seen the inner segments of the bacilla in the Flies, and I described them as a fascelloid
layer f . I did not then understand their significance, and I had been working at the
compound eye for over three years before I prepared a section which enabled me to
recognize these structures as the terminal organs of the optic nerve.

Again, although the structure of the great rods has been very carefully investigated,
that of the neuron has had very little attention paid to it. Even Dr. Grenacher, in his

* L. v. fig. 15. t Li. c. Phil. Trans.

AND THE MOEPHOLOGY OF THE EYE IN INSECTS. 409

elaborate work, only represents portions of it in six figures ; and these are all diagram-
matic outlines, -with but little detail.

M. Berger*, who has given the best figures and descriptions, has only worked with
comparatively thick sections with low powers, and has not only overlooked the bacilla,
but many other very obvious details.

On the otber hand, it must be confessed that many observers have believed that tbey
have traced nerve-fibres from the optic nerve into the great rods or their spindles. Dr.
Grenadier has given one figure, and only one figure, in which he shows this relation.
It represents three of the segments of the dioptron of a Crane-fly t ; and in one of the
three he has shown a nerve passing through the cuticular membrana basilaris, and
ending in one of the cells of the spindle. I have occasionally observed an appearance
similar to that represented in this figure, and my observations have led me to the con-
clusion that the fibres represented are not nervous, but form a portion of the connective
framework. In Tipula the bacilla beneath each segment of the dioptron are enclosed in
a distinct sheath (fig. 57), which often appears to be continuous with the sheath of the
great rod. This gives rise to an apparent continuity between the neural and dioptric
structures in many sections, and may have led Dr. Grenadier to believe that he had
actually traced the nerve into the spindle.

There is no other figure in Dr. Grenadier's work which shows any actual passage of a
nerve-fibre into the great rods ; but two figures represent nerve-fibres passing up to, but
not through, the basilar membrane +. These are both from the eyes of Crustaceans, in
many of which the bacilla are very short. Several figures, of the eyes of Crustacea,
actually show the bacilla, although the author has not referred to these in the text. In
all the other figures, either the neuron is unrepresented, or it is only shown in outline
without sufficient detail.

The retinal layer in Arthropods exhibits two very distinct modifications, which I
propose to term the segregate and the continuous retina.

The segregate retina is characterized by having the bacilla arranged in distinct reti- )
nuke, one for each segment of the dioptron. Each retinula is connected with the
ganglion by a distinct fasciculus of nerve-fibres enclosed in a separate pigmented sheath
(fig. 57). This form of retina is very frequent in larval insects ; it is less often seen in
the perfect state. The Crane-flies and some beetles have a segregate retina in the perfect
condition : Telepliorus affords a good example.

The continuous retina consists of a bacillary layer extending over the inner surface of
the basilar membrane, connected with the ganglion by a single large nerve-trunk,
the fibres of which exhibit a complete decussation (figs. 41, 51, rt). The bacilla always,
however, show a tendency to be collected in small bundles, one corresponding to each
segment of the dioptron, except in some of the Dragon-flies, wherein the central portion
of the bacillary layer, at least, does not exhibit any division into separate fasciculi.

The continuous retina is characteristic of the true Flies (Brachycerous Diptera), of some
Hymenoptera, Lepidoptera, and Neuroptera.

* E. Berger, " Untersuch. u. den Bau des Gehirns und der E.etina des Arthropoden," Arb. Zool. Inst. Wien; torn. i.
(1878). t L. c. fig. 44. i L. c. figs. 109, 114.

410 ME. B. THOMPSON LOWNE ON THE COMPOUND VISION

In many perfect insects, and in some stages of development in others which in the
perfect form bave a continuous retina, conditions exist which are intermediate. For
instance, the bacillary layer may be continuous, but the optic nerve split into numerous
separate bundles ; or the segregate retina may consist of fasciculi so closely united that
it is difficult to decide to which form the modification should be assigned. These varia-
tions are of developmental significance, as will be shown hereafter.

The fibres of the optic nerve are medullated, although in most preparations all traces
of the medullary sheath are lost. It is not, however, difficult to make out the sheath in
recent specimens fixed with osmic acid. I now consider the varicose appearance of the
axis cylinders which I formerly described to be due to post-mortem changes * .

Tbe arrangement of the fibres in separate fasciculi in some insects has aheady been
alluded to. When they form a single nerve-trunk most sections exhibit the complete
decussation of the nerve-fibres. In some planes, however, tbe only indication of such an
arrangement is the cut ends of many of the nerve-fibres. Such variations are undoubtedly
due to tbe plane of the section.

When the retina is segregate, the nervous bundles, which unite the retina with the
ganglion, do not decussate ; but in this case a deeper layer of decussating fibres can
usually be seen in the substance of the ganglion itself (fig. 44).

Supporting fibres connect the inner surface of the retina with the outer surface of the
ganglion, when the nerve consists of a single trunk (fig. 56, cl).

The retinal ganglion is spread over the surface of the optic lobe (oj).) of the cephalic
ganglion (figs. 48 & 49).

It usually exhibits at least four layers, although the number is sometimes greater.

In the Ply (fig. 56) the outer layer (f) consists of small round nuclei imbedded in
a finely granular matrix. Tbe matrix is permeated by fibres connected with a layer
of stellate cells (g), which lie in the central zone of the next layer.

The second layer, beside containing the stellate corpuscles just alluded to, is chiefly
made up of large fusiform cells (cl1), which are connected at both their extremities with
fine fibres. The third layer of the ganglion consists of a number of very fine fibres
(//■), which run parallel to the surface of the ganglion. The supporting fibres of the first
and second layers spread out into foot-like disks on the surface of the fibrous layer (i).
The fourth layer (cl2) resembles the second, except that I have not detected any stellate
corpuscles in connexion with its supporting fibres.

The retinal ganglion is connected with the deeper portions of the optic lobes by fibres
(no2), between which a large number of tracheal tubes lie. These supply the fine tracheal
vessels of the ganglion itself.

In the ganglion of the retina in Agrion (fig. 42) there are two extra layers within the
inner cellular layer — a second layer of fibres parallel with the surface of the ganglion (f2),
and a second nuclear layer (ii2) ; and the ganglion is connected with the deeper structures
of the nervous system by a second layer of decussating fibres (x).

In Tipula and the Gnats the structure of the ganglion is not so complex. In these

* Phil. Trans. I. c. p. 585.

AND THE MOKPHOLOGY OF THE EYE IN INSECTS. 411

insects (fig. 57) I have been able to make out only two layers — a nuclear layer (nuc) and
i layer of large fusiform cells (cl). The latter are connected by a number of nearly parallel
aerve-fibres {no2) with a deep ganglionic layer, which consists of small stellate cells (g).

III. On the Development of the Compound Eye.

The dioptron and neuron are developed from two distinct sources : the former origi-
nates from the hypodermis, the latter from a solid outgrowth of the cephalic ganglia ;
50 far, therefore, there is ground for a morphological comparison between the nervous
and dioptric structures of the vertebrate and compound eye. The dioptron is comparable
with the crystalline lens, whilst the neuron, so far at least, is homologous to the retina.

The researches of Dr. Weismann * have shown that the dioptron in the Fly is formed
from a single layer of cells, and my own observations verify this in the case of the
Lepidoptera aud the Dragon-flies.

It is well known that the outer facets of the eye in the Crustacea are developed later
than the more central facets, the eye increasing in magnitude with each successive
ecdysis. But I do not know that it has been remarked that the peripheral portion of
the eye in insects is less developed than the more central part. I have found this to be
the case even in the fully formed imago, whilst a section taken through the eye of an
immature imago often throws much light on the manner in which the eye is developed.

In larvae in which the compound eye is functionally active such sections are still more
instructive.

Perhaps it is by the gradual addition of fresh facets that the segregate retina takes its
origin. In the Dragon-flies, at least, the continuous retina of the imago is not fully
formed until after the final ecdysis, when it gradually replaces the segregate retina of
the larva.

The dioptron in its earliest stage of development consists of a single layer of columnar
cells, which cannot be distinguished from the ordinary hypodermis of the insect. I have
observed this condition in various Lepidopterous larvse. The columuar cells lie imme-
diately beneath a continuous non-facetted cornea.

At this stage numerous tracheal vessels and stellate cells lie beneath the columnar
layer, and processes from the columnar cells pass into the deeper layer, where they com-
municate with the stellate cells, which belong to the ordinary connective tissue of the
insect.

In this respect the columnar cells closely resemble the cells of the sensory epithelium
in the Medusae and Mollusca. The researches of the brothers Hertwig f show that in
the MedusaB the cells of the sensory epithelium, in the region of the rudimentary eye-
spot, become the terminal organs of the nerve. This may possibly be regarded as au
indication in favour of the hitherto accepted view that the sensory termi nations of the
nerve in the Arthropod are developed from the hypoderm. Such a conclusion is, how-
ever, in my estimation, no more justified than if it were extended to the case of the
crystalline lens of a vertebrate. In the Medusa the epithelial cells, from which the

* Zeitsclir. f. w. Zool. Bd. xiv.

t 0. und R. Hertwig, Das Nervcnsystem und Siunesorgane dos Medusen.

412 ME, B. THOMPSON LOWNE ON THE COMPOUND VISION

nerve-terminals are believed to be developed, are those of an undifferentiated epiblast;
wbilst in the Arthropod, as in the vertebrate, the great neural tract has already beeri
differentiated before the groundwork of the compound eye is laid in the cells of the
hypoderin Metschnikoff*. The hypodermic cells in the region of the dioptron become
greatly elongated, and undergo both longitudinal and transverse subdivision; but as
the tracheal vessels and stellate connective-tissue cells permeate the hypoderin at a verj
early stage in the development of the dioptron, it is extremely difficult to determine how
much of this structure is derived from the hypodcrm and how much originates ii
mesoblastic elements. The earliest stages of development are far more difficult to follow
in the Diptera. This arises from two principal causes : first, the " imaginal disks " froir
which the integumental structures of the imago are developeel are closely united to tht
nervous system ; and secondly, the whole development is very rapid, anel the parts ar<
exceedingly soft and easily destroyed.

The compound eye of the larva of Corethra, in the earliest stage in which I have yei
observed it, closely resembles the same structure (as figured by Metschnikoff t) from Aphii
rosce. It consists of a discoid group of pyriform cells, with their long axes at righl
angles to the surface of the disk. The rounded outer ends of these cells he immediatelj
under the transparent non-facetted cuticle, which represents the cornea, Their innei
extremities are prolonged, and are intimately connected with the deeply pigmented
sheath of the optic nerve.

Dr. Weismann J has figured the disk from a still younger larva, and shows the nerve
ending in a layer of cuboid cells, which, from observations I have made on the very
rudimentary eye of a Lepidopterous larva, form, I believe, a distinct layer beneath the
disk from which the refractive structures are developed. My own observations have
led me to conclude that at this stage the nerve-fibres end in a bed of granular nucleated
protoplasm, not unlike a large motorial end-plate (fig. 73). Each of the pyriform cells
of the disk has a bundle of rod-like bodies of an orange-brown colour near its outer end.
These eventually develop the cone and spindle.

The rods are larger in the cells of the centre of the disk than in those of its periphery ;
and the whole cell shows symptoms of division into four longitudinal parts.

The close relationship of this disk to the cephalic ganglion, both in Aphis and Corethra,
throws light upon the still more remarkable developmental history of the eye in the true
Hies, so well described by Weismann.

In these the disk is attached to the optic ganglion by a narrow pedicle, and has no
connexion with the other integumentary structures in the pupa. It is one of the
remarkable cellular expansions which are known as " imaginal disks." The head and
thorax of the imago are formed from a series of these disks, which are really involutions
of the epiblast. These are drawn into the interior of the embryo whilst it is yet inclosed
in the egg.

The procephalic lobes and all the appendages of the head are distinctly seen as rueli-

* Zeitschr. f. w. Zool. 1871 ; Monthly Mic. Journ. 1872. t Zeitschr. f. w. Zool. Bd. xvi.

t Zeitschr. f. w. Zool. Bd. xiv.

AND THE MOKPHOLOGY OP THE EYE IN INSECTS. 413

nients in the embryo some time before the hatching of the egg ; but before that event
they are withdrawn into the interior of the embryo by a process of invagination during1
the formation of the fore gut, which lies entirely in front of these structures.

The fore gut forms no part of the alimentary canal of the imago, but is shed with the
larval integument in the first stage of the formation of the pupa. This then puts on the
same appearance as the embryo had whilst it still lay inclosed in the egg, by the un-
folding of the imaginal disks ; and these, again, form the anterior part of the body.

After the unfolding of the imaginal disks, the development of the compound eye
proceeds in the same manner as in those insects in which it is developed from the cells
of the hypoderm.

I believe that the pedicle, which supports the disk from which the dioptron is deve-
loped, is merely a portion of the connective-tissue capsule that incloses the ganglion
before the evolution of the disks in the formation of the pupa. The optic nerve and
retina are formed at a later period.

These facts throw light on the relationship between the ordinary compound eye and
the remarkable encapsulated eye of the entomostracous Crustacean, Leptodora hyalina,
described by Dr. Weismann f . This is deeply seated in the median line of the transparent
head, in immediate relation with the cephalic ganglion, and it has no apparent connexion
with the skin or surface of the animal.

In describing the variations of the retina I have already adverted to their develop-
mental significance. I have especially studied the developmental relation of the segre-
gate and continuous forms of retina in the Dragon-flies, and there are numerous indi-
cations that what I have actually observed in these insects is a common phenomenon
amongst the Arthropods. 1\\ the earlier stages of the Dragon-fly larva the compound
eye is already functional, at least as far as its central facets are concerned. In this con-
dition there are very few facets. Those of the more central portion of the eye have a
distinct retinula to each, and each retinula has a distinct nerve-bundle (figs. 43 & 45).
As development progresses, new segments are added to the periphery of the dioptron,
and the number of retinula? increases.

In some sections the continuous retina of the perfect insect (rt ) is already seen upon the
surface of the optic lobe (fig. 44), and the fibres of the nerves from the partial retina? (rt*)
can be seen passing between its rudimentary elements. The existence of a decussating
nerve beneath the continuous retina is very apparent. The retinal ganglion (fig. 44, g)
is now deeply seated in the interior of the optic lobe.

Beneath the ganglion, both in this and in later stages, there is a remarkable ganglionic
centre (marked k in my figures). It is in relation with the retinal ganglion by a
second bundle of decussating fibres. It is, I believe, the representative of those
remarkable kidney-shaped bodies described by Mr. Newton in his admirable description
of the eye of the Lobster % .

As development progresses, the continuous retina gradually enlarges, and approaches

t Zeitschr. f. w. Zool. Bd. xxiv.

t E. F. Newton " On the Eye of the Lobster," Quarterly Journal of Mie. Science, 1873.
SECOND SERIES. — ZOOLOGY, VOL. II. 61

414 MK. B- THOMPSON LOWNE ON THE COMPOUND VISION

the inner surface of the memhrana basilaris of the dioptron (figs. 41 & 45) ; so that at
the final ecdysis it entirely replaces the partial retinae of the larval eye.

This complete change of the larval retina at the final ecclysis is undoubtedly a very
remarkable phenomenon. When, however, the epiblastic nature of the nervous system
is borne in mind, it is not, perhaps, more remarkable that an ecdysis should occur in
relation to the nervous than that it should occur in relation to the cutaneous epithelia.
It has occurred to me that the kidney-shaped ganglia already alluded to, which appa-
rently vary much in number in different stages of development, may be really successive
rudimentary ganglia and retinae destined in turn to become functional.

The extent of the final continuous retina varies much in different species of insects.
In Agrion the whole inner surface of the basilar membrane is in contact with the con-
tinuous retina in the fully formed imago ; but in the great eyes of JEshnla it only replaces
the central retinulse. The same partial replacement of the retinulse is the condition in
many insects in which the eye is only functional in the imago. In these it is probable
that the central portion of the retina is continuous from its first formation, and that it
is, as it were, supplemented by peripheral partial retinre.

Most of the Lepidoptera apparently exhibit this condition, the continuous retina being
most developed in the Diurna. As has been already stated, the larval condition is per-
manent in the Crane-flies and Gnats ; whilst in the Diptera with short antennae, Muscidse
and Tabanidae at least, where the formation of the pupa almost partakes of metagenesis,
no partial retinae are ever formed. It appears as if the developmental processes had
been much abbreviated in these insects, and that the final stage is reached by a single
and complete metamorphosis.

IV. On the Morphology of the Eyes or Arthropods.

Although of late the views of Midler with regard to the relations of the simple and
compound eyes of Arthropods have fallen into discredit, I must return to these views.

The eyes of the Isopod Crustaceans, which he spoke of as aggregate, are undoubtedly
really intermediate between a simple and compound eye. The transition from a few
widely separated simple eyes, which form the aggregate eye, is so natural that it has only
obtained discredit because there is a wide difference in the structure of the stemmata
of insects or Arachnids, and of the compound eye.

The case is, however, otherwise with the simple eyes of larval insects, which are almost
identical with a segment of the dioptron of the compound eye. Landois *, describing
the eyes of several caterpillars in 1866, remarked that they present a condition inter-
mediate between that of the compound eye and the ordinary simple eye, and proposed
the term "ocelli comjjosltl" for them. I have figured sections of the eyes of three
noctuid. caterpillars (figs. 36-40); these exhibit a lens consisting of three segments placed
immediately beneath the very convex corneal facet.

* Zeitschr. f. w. Zool. Bd. xvi.

AND THE MOEPHOLOGT OF THE EYE IN INSECTS. 415

The lens is not unlike the crystalline lens of a vertebrate. It is albuminous, or at
least gives a characteristic yellow with nitric acid.

Beneath the lens a fusiform spindle is seen, very similar to the spindle of the com-
pound eye. This has a retinula beneath it consisting of a small fasciculus of bacilla.
Several of these eyes are united by a common optic nerve, like the stemmata of insects,
and three usually form a group.

I believe that the spindle in this form of eye has the same function as that which I
have assigned to it in the compound eye.

A comparison of the figures given by Dr. Grenadier * of the eyes of Porcellio and of
Sialis larva with those of the ocelli of caterpillars which accompany this paper, will show
how close the relation of these structures is to each other. On the other hand the
absence of a subcorneal lens and spindle, as well as the arrangement of the retinal
elements, in the simple ocelli of insects and arachnids, show that these organs are formed
on a type which differs essentially from that exhibited by the compound eye.

1 believe, however, that the columnar cells immediately beneath the cornea, the
vitreous of Dr. Grenadier, represent the dioptron ; they are undoubtedly of hypodermic
origin, and are separated from the retina by a fibrous membrane which apparently
corresponds to the membrana basilaris of the compound eye. Dr. Grenadier figures
and describes this membrane t in a species of Salticus, and traces the fibres of which it
is composed, from the outer ends of the bacilla of the retina to a number of nuclei
situated in a sinus which surrounds its margin. I have made some remarkably good
sections of the eyes of Salticus scenicns, in which it is quite easy to see that no such
connexion exists between these structures.

In some of these sections the fibrous membrane has completely separated from the
bacilla, just as the membrana basilaris separates from the retina in the compound eye.

The sinus (Is, fig. 33) which surrounds the membrane in Salticus contains radiating
fibres very similar to those which I have described in the sinus around the margin of the
membrana basilaris of the compound eye ; and it is these fibres which contain the nuclei
to which Dr. Grenadier believes he has traced retinal fibres.

At present the origin of the retina of the simple eye cannot be said to have been
determined ; I have sought in vain for any reliable indications as to its origin. Dr.
Grenadier believes it to arise by a modification of the cells of the hypodermj. His
arguments in favour of this origin are very unsatisfactory, and apparently indicate that
the vitreous and not the retinal elements arise from this layer.

A cellular vitreous is always present in the simple eyes of insects : I formerly failed to
demonstrate it in the stemmata in Flies, but I have since found that this layer exists,
although from its extreme thinness it is not easily seen except in specimens preserved
with osmic acid §.

On this point, at least, I must endorse the views of Von Graber ||, rather than those of
Dr. Grenadier. In a very able paper by Prof. E Ray Lankester and Mr. A. G. Bourne ^f,

* L. c. figs. 12, 95, 96, & 97. f L. c. t L. c. § Phil. Trans. 1. c.

II Archiv f. mikrosk. Anat. vol. xvii. 1880, p. 58. If Quarterly Journal of Microscopic Science, Jan. 1883.

Gl*

416 MR. B. THOMPSON LOWNE ON THE COMPOUND VISION

the vitreous is said to be absent in the lateral eyes of Scorpions, although it is present
in the more highly developed central eye. This is a remarkable exception to the
conditions which I have observed ; but as I formerly overlooked the vitreous in cases
in which I now know it exists, I cannot help suspecting that the cells have been destroyed
in soine way in the preparation of the sections.

I propose the following classification of the visual organs of the Arthropoda.

I. Simple ocelli. III. Aggregate eyes.

II. Compound ocelli. I.V. Compound eyes.

I. Simple ocelli. — I include under this term the ocelli of Arachnida aud the
stemmata of perfect insects, and I think it probable that the eyes of the Myriapoda
consist of clusters of such stemmata ; but at present I am not sufficiently acquainted with
the modifications of the eyes of this family to speak with any degree of certainty, as the
investigation of the eyes of Myriapods is exceedingly difficult.

II. Compound ocelli. — I use this term to indicate the ocelli of larval insects in
which there is apparently a second refractive system — the spindle, which magnifies and
erects the subcorneal image.

I have little doubt that the eyes of the Coryceidce, at least those of Cojjilla, belong to
this class ; but I have not had any opportunity of examining them, and only judge from
Dr. Grenadier's description and figures.

III. Aggregate eyes. — I include in this division the semi-compound eyes of Isopods,
which appear to be nothing more than aggregations of compound ocelli.

Dr. Grenadier regards the subcorneal lens of the Isopod as a highly modified crys-
talline cone. I regard it as the representative of the oil-lens of the compound eye, and
of the lens of the compound ocellus ; perhaps the cone of the compound eye should also be
regarded as a highly modified form of lens ; at least it departs more from the primitive
type than the spheroid lens of the Isopod.

IV. Compound eyes. — These are the ordinary eyes of the Crustacea and Insecta.
They exhibit very various forms, and many efforts have been made to classify them,
chiefly founded on the variation of the dioptron, and especially those of the cornea and
crystalline cone.

All these classifications appear to me unsatisfactory from a morphological point of
view, as they do not harmonize with the affinities of the forms of the Arthropods in
which the variations occur, neither do they throw any light on the genetic relations of
the compound eye.

Perhaps the following classification will be useful in making the relations of the
various forms more easily comprehended.

1 am so little acquainted with the structure of the eye in the Crustacea from personal
observation that I shall confine my remarks chiefly to the conditions which I have
observed in the Insecta.

(I.) Eyes in which the dioptron is incompletely separated from the neuron, each
segment of the former having a distinct retinula. I have observed this condition in the
larva of the gnats, in Tipula and in the genus Telephorus amongst the Coleoptera.

AND THE MORPHOLOGY OP THE EYE IN INSECTS. 417

(II.) Eyes in which the dioptron is completely separated from the neuron by the
membrana basilaris, but in which the retina is divided info distinct retinulce. This is
the condition of the compound eye of the Dragon-fly larva, in the imago of the Gnats,
and in the Orthoptera.

(III.) Eyes in which the dioptron is completely separated from the neuron, and the
retina is continuous, but in which the nerve-fibres of the optic nerve are arranged in
distinct fasciculi : this condition holds in many Hymenoptera, Hemiptera, and Lepidoptera.

(IV.) Eyes in which the dioptron is completely separated from the nem*on, in which
the retina is continuous, and the fibres of the optic nerve form a single bundle and
decussate completely . This condition is found in the Muscidee, the Diurnal Lepidoptera,
and in the perfect form of some Dragon-flies {Agrion).

It would appear therefore that the compound eye is to be regarded as a collection of
compound ocelli, in which the tendency to close union is greater nearer to the surface
than in its deeper portion, the optic nerves being the last parts to become fused into a
single compound structure.

DESCRIPTION OF THE PLATES.

The small letters indicate the following parts in all the figures.

i. Iris cells.

up. Optic gauglion.

k. Deep or kidney-shaped gan-

pg. Pigment.

glion.

/;/'. Radating fibres.

/. Lens.

/•/. Retina.

Is. Lymph-sinus.

//-.Partial Retina.

in. Basilar membrane.

sh. Sheath of the great rods

mc. Ciliary muscle.

sp. Spindles.

re1. re2. Nuclear layers.

sr. Scleral riug.

nuc. Nuclear bodies.

tv. Tracheal vessels.

no. Optic nerve.

v. Vitreous humour.

Plate XL.

c. The cornea.

cc. The crystaline cone.

en. Subcorneal nuclei.

eh. The chamber.

c6. Ciliary bodies.

cll.cP. Outer and inner cellular
layers.

gx.g2. Sections of the optic gan-
glion.

h. Hypodermis.

Fig. 1. A semidiagrammatic section of the eye and optic ganglion of a Blow-fly.

1 a. The membrana basilaris of the same insect seen from its neural surface, ped, pedicle of the scleral
ring. Both figures show the large muscle m.c. attached to the inner edge of the scleral ring, sr.

2. A diagram showing the manner in which the image is formed on the Arthropod retina.

3. An optical section of two of the segments of the dioptron of the Plume Moth (Pterophorus pen-

tadactylus) .

4. A diagram showing the optical relations of the same.

5. A semidiagrammatic representation of a vertical section through the great eyes of a Dragon-fly

(JEshnia), showing the lymph-channels ; a.a. afferent vessels; b.b. efferent openings.

6. A section through the scleral ring and membrana basilaris in the region of the efferent lymph-

path, showing the radiating fibres )/, from the eye of yEshnia.

7. A similar section from the eye of Agrion.

8. A section through the scleral ring, s.r., of Agrion, showing a marginal lymph-opening.

9. The afferent vessels of the dioptron of sEshnia.

418 MR. B. THOMPSON LOWNE ON THE COMPOUND VISION

Fig. 10. An optical section of a segment of the dioptron of Tipula in the recent condition.

11. The spindle of the eye of the same, slightly altered by the process of preparation.

12. A similar spindle.

13. A segment or tube of a spindle from the same insect, showing the contracted tnbe with soft

vacuolated material adhering to it, possibly a portion of its contents.

14. a. A transverse section through the spindle of the same insect.

14. b. A similar section, in which the peripheral tubes have partially collapsed.

15. Three of the peripheral segments of the dioptron of a Blow-fly in the immature condition.

16 to 18. Portions of the great rods of a Blow-fly, showing the different appearances which they
present, owing to partial or total collapse of the spindle.

16. A portion of the spindle partially emptied by pressure.

17. A specimen showing still further collapse of the tubes.

18. A portion of a spindle showing vacuolation of the sheathing-eells and collapse of the tubes in

normal saline solution.

19. The lens at the outer end of the spiitdle from the same insect, from an osmic-acid preparation ;

rate, spheroids in the interior of the lens, sp. spindle.

20. A transverse section of the great rods, spindles, sp., sheathing-cells sh., and tracheal vessels, tv.,

from the same.

21. The lcus-capsule, ruptured by pressure, from the Earwig.

22. A single segment of the dioptron, and a few bacilla from the Yellow TJnderwing Moth (Triphcena
pronuba), from a specimen prepared in chloral hydrate solution.

Plate XLI.

Fig. 23. A section through the entire eye of a Noctuid (Xy/ophasia po/yodon).

24. Transverse sections through the cone of the same insect : a, near the base ; b, near the apex,

showing the coloured granules.

25. Transverse sections through the rods and spindles of the same insect: a, b, c, three successive

sections through the spindle-sheath ; d, section through the spindle showing the cells of the
sheath ; e, section through a stellate spindle.

26. The chamber and cone of the same insect. The tube has collapsed.

27. The tube between the cone and spindle, from the eye of a species of Noctuid Moth ; the tube

contains unicellular parasitic organisms, z.

28. The chambers and a portion of the spindles of two segments of the dioptron, an osmic-acid

preparation from the eye of a Blow-fly : oc, outer portion of the cone ; cc , crystalline cone.

29. The spindle from the same, mounted in a dilute solution of osmic acid.

30. A section of the cornea of an African Carpenter Bee (Xylocopa), from a dried insect, showing
its laminated structure : c1, lenticular layer ; c2, laminated layer ; c3, inner surface.

31. A single segment of the dioptron of an Ant (Formica rufa) , showing the ciliary rods cb. ; 31 a, one

of the ciliary rods detached and more highly magnified.

32. A similar preparation from the eye of a Wasp ( Vespa germanica) : 32 a, one of the same detached

and more highly magnified ; cb1, refractive granule ; cP, iris cell ; cb3, fringe-like process.

33. A portion of the eye of a Spider (Salticus scenicus), showing the fibrous membraue m, which

separates the dioptric from the nervous structures ; Is, lymph-sinus, containing radiating
nucleated fibres.

34. Two of the eyes of Salticus scenicus.

35. The simple ocellus of a Blow-fly : 35 a, one of the retinal elements more highly magnified ;
r1, outer refractive segment ; r2, intermediate pigmented portion ; r3, nucleated protoplasmic
portion.

AND THE MORPHOLOGY OF THE EYE IN INSECTS. 419

Fig. 36. One of the confound ocelli of a Noctuid caterpillar : h, hypoderni ; py, pigment covering the
spindle.

37. One of the compound ocelli from another Noctuid caterpillar.

38. A similar preparation from a third species of Noctuid caterpillar : no, union of three nerve-
trunks from three compound ocelli.

39. An oblique section through a similar eye in an embryonic condition.

40. A transverse section of a compound ocellus from a Noctuid caterpillar, showing the iris and lens

in situ.

Plate XLII.

Fig. 41. A transverse section through the entire eye of a Dragon-fly (Ayrion viryo), from the perfect
insect.

42. A highly magnified representation of a portion of the retinal ganglion, from the same prepa-
ration : iil, nuclear layer; clx, cellular layer; f1, fibrous layer; cl2, second cellular layer; f2,
second fibrous layer ; n2, second nuclear layer ; x, second decussating layer of fibres.

43. A section through the entire eye of a larval Ayrion, from a specimen 6 lines long : rt*, segregate
retina; yl, ganglion.

44. A similaiv section from a more advanced larva.

45. A similar section from an adult larva.

46. The bacilla from the neuron of the same larva.

47. The bacilla from the neuron of a Dragon-fly larva {Libelhda depressa).

48. The cephalic ganglia of a Cockroach, from a large larva, seen from behind : op, optic lobe ;

n, nerves to partial retinae.

49. The same, seen from before.

50. A transverse section through one of the partial retime in a rudimentary stage : no, one of the

optic uerves.

51. A section through the neuron of a Fly (Syrphus ribesii).

52. A portion of the retinal ganglion of the same, seen with an immersion lens j^. (The references
as in fig. 42.)

53. Two bundles of bacilla from the retina of the Blow-fly, from an osmic-acid preparation, showing

the tracheal vessels and cells of the neuroglia : o outer and i inner segments.

54. A transverse section through the entire eye of a Gnat.

55. Bacilla from the eye of Tipula, showing the axis-cylinder between the inner and the outer

segments.

Plate XLIII.

Fig. 56. A semidiagrammatic representation of the retina of the Blow-fly, drawn from preparations
prepared in different fluids, and showing the connexions of the various elements : a, neuroglia,
of the retina, from a chloral hydrate preparation ; b, the same, from an osmic-acid preparation ;
" c, the bacilla, from osmic-acid preparations ; B1, outer segments ; B2, inner segments ; d, the
neurogba of the optic nerve ; e, the fibres of the optic nerve, from osmic-acid preparation ;
/, outer layer of ganglion, osmic acid ; g, cells with the supporting neuroglia, chloral hydrate ;
h, transverse fibres, Midler's fluid; »r, inner limiting membrane of retina ; m3, endothelium
covering the ganglion ; no', inner nerve-fibres connecting the optic and deep ganglion. The
other references are the same as those in the rest of the figures.
57. A portion of the neuron of a Crane-fly {Tipula oleracea). The references as in fig. 56.

420 COMPOUND VISION AND MORPHOLOGY OF EYE IN INSECTS.

Fig. 58. A portion of the retina of a Blow-fly, from a specimen prepared with chloral hydrate, and
stained with eosin.

59. A portion of the retina of the Blow-fly, osmic acid.

60. A portion of the retina of a Blow-fly, Midler's fluid.

61. A transverse section of a portion of the retina of a Blow-fly, from a specimen prepared with
Midler's fluid and mounted in balsam.

62. A transverse section of a portion of the retina of a Blow-fly, through the inner segments of the

bacilla, from a specimen preserved in chloral hydrate and subsequently mounted in balsam.

63. Two fasciculi of bacilla, from a Hawk-Moth pupa.

64. Bacilla of a Blow-fly, showing the lenticulus, from an osmic-acid preparation.

65. Bacilla from the eye of a Cockroach, with double outer segments.

66. Similar bacilla, from the eye of a Blow-fly.

67. Bacilla from the eye of a Noctuid Moth.

68. Tracheal vessels on the inner surface of the membrana basilaris of the Blow-fly.

69. Retinal pigment-cells and a portion of the basilar membrane of a Dragon-fly (JEshnia grandis).

70. A portion of the basilar membrane from the eye of a Lobster, showing prismatic thickenings
at the inner extremities of the great rods.

71. A portion of the retina and basdar membrane of the Cabbage Butterfly, showing lenticular
thickenings.

72. A vertical section through the basilar membrane of Syrphus, showing the cellular sheaths of
the fasciculi of bacilla, and the cells around the necks of the expanded tracheal sacs of the
dioptron.

73. The disk from which the retina originates, from a Lepidopterous larva (Noctuid) .

74. Two segments of the dioptron, with their retiuulae, from the eye of a perfect Gnat {Corethra
plumicornis) : 74 a. A transverse section through the inner extremity of the cone of the same ;
74 b. A transverse section through the spindle of the same.

75. A portion of the dioptron of the larva of the same insect.

76. A section of a portion of the eye of a young Gnat larva.

77. A portion of the dioptron of an immature Crane-fly (Tipula oleracea), which had just escaped
from the pupa.

Note. — Since the above was sent to press, my attention has been drawn to a short
paper by Justus Carriere of Strassburg (Quart. Journ. of Micros. Sci. Oct. 1884, p. 673),
" On the Eyes of some Invertebrata," in which my retinal layer is figured. He speaks
of it as the " palisade layer." He remains, however, a disciple of established views, and
has not given the retinal layer nearly so much attention as it deserves. So far as his
observations go, they appear to me to confirm my own, as I firmly believe that
M. Carriere, on further investigation, will be led to admit the validity of my views.

r 421 ]

XV. Contributions to the Knowledge of the Genus Anaphe, Walker, By Lord

Walsingham, M.A., E.L.S.

(Plates XLIV., XLV.)

Read March 20, 1884.

£ OB the specimens described in the present communication, I am indebted to the kind-
ness of Colonel J. H. Bowker, of Durban, Natal. In a letter dated July 4th, 1883, he
wrote to me as follows : — " I am sending by the ' Moor,' Union steamship, a little box
with a nest of the congregating moth. The larvae are most interesting, often denuding a
tree of its foliage ; they move in a body, sometimes ten or even twenty yards long, in
search of ' pastures new,' and when the time comes, form into a cluster, and form the nest
covered with a brown silk. Unless the change of climate has an effect, they will come out
about September next. I should suggest the nest being set up in a green-house, not too
hot, with a view towards the morning sun. A small nest which I sent to the Cape Town
Museum produced about 80; judging from the size of the two, the one I send ought to
produce about 300. The natives use the silk for medical purposes, somewhat as we use
oiled silk." The box reached me at the beginning of August 1883, and I was much
surprised to find that the larvae were alive and apparently healthy. Many of them
remained in the so-called nest, but bodies of from twenty to forty constantly came out
and moved about, always keeping close to it. They moved in a closely packed mass,
following a slightly curved line, much in the same position as that in which the small
cocoons may now be seen arranged in the interior of the large one.

On the 9th of August I sent them to the Insectarium in the Zoological Society's
Gardens, in Regent's Park, where they have been under the attentive care of Mr.
Arthur Thomson until about the middle of March. Mr. Thomson informs me that about
250 moths emerged from the large cocoon. The first on December 3rd, the last on
February 11th. One pair copulated and produced eggs. The eggs hatched, but although
the name of the food plant had been ascertained from Mr. G. Baker, at Kew, to be
Bridelia micrantha, Baillon, the dead leaves attached to the nest having enabled him to
identify it, it was found impossible to obtain, there or elsewhere, any thing sufficiently
nearly allied to it to induce the young larva? to feed, and they soon all died. Mr. Thomson
was successful, at my request, in finding in the glass case a few of their bodies, much
shrivelled, but, nevertheless, most interesting, as indicating the probable affinities of the
genus to which they belong. The appearance of the moths was followed by the emergence
of a number of dipterous parasites. The larvae did not finally enclose themselves in the
large cocoon until about a fortnight after they reached the Insectarium ; they must
therefore have been about 50 days without food. Some of them died, probably from
this cause, without entering the cocoon at all ; and I find that one or more have spun
themselves up in the outer covering of the main cocoon, without coining to maturity.
It will be seen that Colonel Bowker's estimate of the number sent was remarkably correct.

SECOND SEKIES. — ZOOLOGY, VOL. II. 02

422 LOKI) WALSINGHAM ON THE GENUS ANAPHE.

In the British Museum are at least four described species of the genus Anaphe, viz . —
Anaphe venata, Butler, from Old Calabar ; Anaphe ambrizia, Butler, a small species
from Angola ; Anaphe reticulata, Walker, from Natal ; and Anaphe panda, Boisduval.

The single typical specimen of Anaphe panda, from Natal, is precisely similar to those
from Colonel Bowker ; but placed under the same name is a considerable series of specimens
which differ from these in the absence of any transverse dark bar crossing the pale trian-
gular or wedge-shaped space beyond the middle of the anterior wings. In those specimens
that space is clear and unclouded, the longitudinal brown lines being confined to the
outer side of the oblique fascia, and connecting it, as in the Natal form, with the dark
fringes of the apical margin. The brown markings are also somewhat darker than those
of A. panda. These specimens are from Monga-ma-Lobah, in the Cameroons, and
although probably only a local race, arc at least as much entitled to specific distinction
as is the species named by Walker, Anaphe reticulata. The entire absence of variation
throughout the large colony sent by Colonel Bowker seems to indicate that the slight
differences observable in these local races are constant and reliable, and that there is
little or no individual variation.

Anaphe infracta, sp. n. (PI. XLV. fig. 8).

Head and palpi ferruginous ; antennae black. Thorax creamy white anteriorly,
ferruginous posteriorly, the white divided by a ferruginous streak along the middle,
reaching to the head. Fore wings creamy white, with narrow brown margins, the
fringes also brown, two narrow brown transverse bands ; the first, arising before the
middle of the costal margin, tends obliquely outwards to a point slightly beyond the
middle of the dorsal margin, where it joins the lower extremity of another narrow brown
band, which runs to the costa parallel with the apical margin, the two enclosing a plain
wedge-like space of the pale ground-colour of the wing. The outer of these bands is
connected with the apical margin by two shorter brown streaks from above and below
its middle, the upper one slightly depressed outwardly, the lower one more depressed,
reaching the margin slightly above the anal angle. Ilind wings creamy white, tinged
with ferruginous at the base, and about the abdominal margin, with a very faint indica-
tion of a transverse median shade of the same colour, more visible on the under side.
Abdomen pale ferruginous, with darker lines at the junction of the segments.

Expanse of male 47 mm. ; of female 58 mm.

Monga-ma-Lobah, Cameroons ( G. Thomson).

Anaphe reticulata, Walker, figured by Herrich-Schaffer (Sammlung ausser-europai-
scher Schmetterlinge, f. 434), under the name Arctiomorpha euprepiarformis, has two
distinct transverse bars across the wedge-shaped pale space, the lower one of the two
being continued to the base of the dorsal margin. Two cocoons, said to belong to this
species, are in the British Museum, and differ from the one now under notice, so far as
I can ascertain, only in their rather smaller size and more irregular shape, and in the
colour of the silk of which the small cocoons in their interior are composed. These
(which in my specimen arc white, with scarcely any brownish tinge) are, in those cases,

LORD WALSINGHAM ON THE GENUS ANAPHE. 423

f a rich brown, almost exactly the same colour as the external covering of the aggregate

lass

A sixth species was kindly given to me about three years ago by Mr. G. T. Carter,
ollected by him at Aburi, od the Gold Coast. The single specimen (a male) is now in
lie British Museum.

The colouring is precisely similar to that of A. panda. It differs from all the described
pecies of this genus in having a single band only, crossing the anterior wings ; moreover,
his band, instead of being somewhat curved outwards, as is the first bar iu all other
nown species, is curved slightly inwards towards the base. The costal margin is brown
hroughout ; the dorsal margin also brown, except a very short space near the base.
Tie fringes of the anterior and posterior wings are also brown. On the underside the
ipper portion of the central band of the fore wings is visible, and there is a distinct
:ostal spot on the hind wing, which is not visible above.

Anaphe Carteui, sp. n. (PL XLV. fig. 9.)

Face and palpi ferruginous; head creamy white above; antenna} black. Thorax
jreamy white in front and at the sides, with a ferruginous blotch behind projecting
toward in the middle, in an obtuse wedge-shaped form, but not reaching to the head.
Pore wings creamy white, the margins narrowly ferruginous, with a narrow transverse
median band of the same colour, which is slightly bent inwards ; the fringes also
ferruginous. Hind wings creamy white, with ferruginous fringes, except on the
abdominal margin.

Expanse 36 mm.

Aburi, Gold Coast {Carter).

I find in the Proceedings of the Entomological Society, June 5th, 1878 :—

« Mr D Grei- Eutherford exhibited a series of large cocoons sent by Mr. George
Thomson from Mount Camaroons, West Africa, formed by the caterpillars of a species
of Bombyx allied to Anaphe panda, Bdv. They were taken at an elevation of 5000
feet above the level of the sea, and, judging from the nature of certain twigs and pieces
of native cord fastened to their exterior covering, appeared to have been found attached
to fences or to the eaves of houses. The cocoons are light brown in colour, and very
irregular in shape, rather flatfish, and vary in diameter from four to seven inches.

"Each cocoon contains from 130 to 150 separate cocoons, each enclosing a larva or a
chrysalis in various stages of development, the least mature being near the centre of the
aggregate mass. To some of the cocoons were attached a case containing what appeared
to be the larvse of some species of Ichneumon or dipterous parasite.* All the smaller
cocoons are firmly bound together by layers of a coarse silken material, this aggregation
bein- enveloped by a thin brown lining almost of the consistency of parchment, the
whofe forming a compact mass, which again is surrounded by a loose silken tissue, while
a stron- covering of the same material envelops the whole. Erom this it would appear
that A panda like some other species of Bombycidce, is social, and that the caterpillars
unite in considerable numbers in order to form an aggregate cocoon of sufficient strength

* These were subsequently proved to be Hyinenoptera, sp. Oryptus formosus, Bridle.

62*

424 LORD WALSINGHAM ON THE GENUS ANAPHE.

to preserve the chrysalis from the attacks of enemies, and probably also against extreme
changes of temperature."

It may be noticed that this description applies in almost every detail to the specimens
under consideration, but I would point out one exception.

The plan adopted by the numerous members of this South-African insect republic
appears to have been based upon a more perfect recognition of the equal rights of indi-
viduals. There is no necessity under their system to trust to the retarded development
of those situated towards the interior of their habitation. There are two outlets, and the
inhabitants have ranged themselves in two almost equal bodies in such a manner as to
secure to each moth a ready means of egress from one or other of these openings. The
points of the numerous cocoons are all directed as nearly as possible towards the nearest
available outlet, and a separate silken passage seems to have been prepared for the ac-
commodation of each occupant, so that the foremost ones in no way block the way of those
situated behind them. Some few of the moths which appear to have been unable to get
out, are probably those of which the larvie were somewhat starved, and therefore imper-
fectly developed, and which were thus not upon equal terms with their more advanced
companions. There are probably some Fellows of the Society who are better capable of
judging of the quality of the silk than myself. It seems to me to be extremely
strong and tough, although, except on the external covering of the colony, it is not
particularly coarse. On winding a small quantity, it appeared to be perfectly con-
tinuous, but not easily run off. Although there is far less in quantity on each
separate cocoon than on those of Bombyx mori and other silk-producing Bombyces, the
immense number of cocoons in each colony might not impossibly render it of some
economic and commercial value. A silk-plantation on the high lands of Natal would
certainly be a pleasant and healthful field of enterprise, and might perhaps turn out to
be a profitable one.

The only description I have been able to find of a larva of the genus Anaphe is given
by Carl Fromholz in a paper in the ' Berliner entomologische Zeitschrift,' Bd. xvii.,
1883, Heft i. pp. 9-13, which is illustrated by an octavo plate. This larva was supposed
to be that of Anaphe panda, and the large cocoon figured on the plate was evidently
made by this or some allied species, but the specimens came from the interior of Africa,
250 miles from Dara Salam, and the moth does not appear to have been actually identi-
fied. This larva was described from the dead, and somewhat shrivelled specimens which
reached the author, as " yellowish white," which by no means agrees with the colour of
those now exhibited.

The examples now figured were preserved by a process which I have successfully
applied to those of the majority of the British Macro-Lepidoptera, as well as to numerous
exotic species. They have the head black, divided by a narrow pale line along the middle ;
the second segment, with two narrow transverse horny plates, separated from each
other in the middle, and a narrow black horny plate on the anal segment. The body is
of a purplish vinous colour, slightly paler between the segments, witli a faiutly indicated
pale line above the black spiracles, and some pale yellowish white spots at the junction
of the segments below them.

LORD WALSINGHAU ON THE GENUS ANAPHE. 425

The whole body is covered with short black hairs. On the upper portion of each
segment are two pale wart-like spots, situated to the right and left of the dorsal vessel.
Behind these, and more widely separated from each other, are two smaller, but other-
wise similar., spots, whitish hairs springing from all four. Other wart-like spots less
distinctly paler than the surrounding skin are arranged on the sides of the segments in
the following order ; one above and one below, and a third behind and slightly below
each spiracle. There is also one above each of the abdominal claspers, which are dull
yellowish white, with black booklets, as are also the anal claspers. The prolegs are black,
with a whitish ring around the base of each. All the hairs have small, but sharp spines
around them, arranged somewhat in a spiral form from base to tip, the points directed
outwards. This formation, as in the case of those of many larva? of the Lasiocampidse
and Liparida?, notably in that of Cnethocampa pityocampa, "W.V., renders them very
irritating and difficult to eradicate if, by any means after detachment, they become
lodged, with the base downwards, in the human skin.

This genus has been placed among the Liparidse. In the Annals and Magazine of
Natural History (4), xix. p. 462, Mr. Butler assigns to it a position between Marana and
Numenes, and suggests that A. reticulata and A. panda are probably varieties of one species.

The young larvae above described have so much the character of Arctiidae, with their
clothing of long brown erect hairs, as to suggest the idea that the genus should more
properly be included in this family. The most immature forms of larvae may usually
be relied upon to indicate their affinities and probable derivation. In this case it is a
fair inference from the evidence they afford that Arctia is an older form than Liparis,
and that the genus Anaphe occupies the position of a connecting link between the two
families.

Herrich-Schaffer, in creating for this form the generic name Arctiomorpha, showed
that he had observed its peculiar resemblance to Arctia ; but he placed it in the family
Notodontidse.

Boisduval, in describing this same species in the ' Voyage de Delegorgue en Afrique
Australe,' torn. ii. p. 600, 1847, mentions its resemblance to Arctia villica, which I
confess I am unable to recognize. The large anal tuft of easily detached hairs on the
body of the female is the point in which it most strongly resembles the typical forms of
the genus Liparis, as well as in its neuration, the median vein of the hind wings having
but three branches instead of four as in the Arctiida?.

I am informed by Mr. Kirby that the dipterous parasites present in the cocoon from
Colonel Bowker appear to be closely allied to Tachina onchcstus, Walker (List Dip. B.
M. iv. p. 773, 1849). [Two of these emerged on the day the paper was read, and one was
shown alive at the Meeting of the Society, along with the series of specimens described
in this communication.]

Fromholz, in the paper already referred to, describes under a new genus of Lepidoptera
belonging to the Phycida?, Zophodiopsis hyceiiella, as infesting nests of the species
mentioned by him and believed to be Anaphe panda. We have therefore the Lepido-
ptera, Hymenoptera, and Diptera, each represented as pax-asitically attached to the larva;
or to the habitations of laiwas of this interesting genus.

SECOND SERIES. — ZOOLOGY, VOL. II. 63

426 LORD WALSINGHAM ON THE GENUS ANAPHE.

EXPLANATION OF PLATES.
Plate XLIV.

Fig. 1. Enveloping cocoon of Anaphe panda, Boisd., two-thirds natural size — outside view.
Fig. 2. The same cut open, showing internal texture and arrangement of small separate cocoons. o\ ol,
twigs inserted into the openings.

Plate XLV.

Fig. 1. Portion of inner surface of enveloping cocoon, showing lining between inner wall and exterior,
with modification of smooth surface by introduction of foreign body. 1 a. Fibre of envelope,
highly magnified.

Fig. 2. Diagram of cross sectiou through the middle. 2 a. Outlet at back of same.

Fig. 3. Full-grown larva. 3 a. Spines of larva, highly magnified.

Fig. 4. Young larvae.

Fig. 5. Separate cocoons from interior. 5 a. Empty pupa-shell.

Fig. 6. Single brown cocoon, from interior of similar colony in the British Museum.

Fig. 7. Anaphe panda, Boisd. £ .

Fig. 8. infracta, Wlsm. c? .

Fig. 9. Carteri, Wlsm. $ .

Fig. 10. Dipterous parasite, Tachina onchestus, Walk. 10 a. Empty cocoon of same.

[ 427 ]

XVI. On a new Species of Ccelacanthus (C. Tingleyensis)/ro?ra the Yorkshire Cannel
Coal. By James W. Davis, F.L.S., F.G.S.

(Plates XLVI.-XLIX.)

Read Juno 19th, 1884.

SPECIMENS vary very much in size, and occur in all stages of growth between
3 inches and six times that length ; their form is somewhat long and slender, with a
series of powerful fins well adapted for rapid progression through the water. Examples
which appear to have attained the maximum size of ahout 18 inches in length are taken
for the following descriptions. In such a one the head occupies one fourth the entire
length, the body, from the pectoral arch to the base of the tail, two fourths, and the tail
the remaining fourth. The greatest depth of the body is immediately behind the pectoral
fins, where it is 35 inches ; posteriorly the depth diminishes gradually to the base of the
tail, which is 2 inches. The hody was probably more or less cylindrical in form, and the
lateral diameter little less than that between the dorsal and ventral surfaces. The
anterior portion of the caudal fin expands to a diameter of 4 inches, and the rays radiate
from the base 3 inches : beyond this the second portion of the tail extends about 1*3 inch,
its termination supporting a second series of rays, forming the peculiar prolongation of
the caudal appendage characteristic of the genus. In addition to the caudal tin there are
seven others — two dorsal and one anal fin, a pair of pectorals, and a pair of ventrals. The
posterior dorsal fin (PL XLVI. d2) is located 2-5 inches in front of the caudal (c1), and is
supported by a scale-invested lobe from the body. The anterior dorsal fin (dl) is about
3 inches in front of the posterior one : in several specimens where this fin is present
there does not appear to be any lobe supporting it. The pectoral fins (p) are situated
immediately behind the thoracic arch, and, like the ventrals (v) and anal («), they are
lobate. The ventrals are powerful and large fins, placed opposite the space intermediate
between the two dorsal fins. The anal fin, 3 inches behind the ventrals, is situated at
a considerably shorter distance from the caudal than is the posterior dorsal fin. The
whole of the body is invested in a covering of comparatively thin, enamelled, and
beautifully ornamented scales ; those on the dorsal and lateral regions are broader
and have a more obtuse termination than those on the ventral surface. The scales
covering the lobes of the fins are smaller than those of the body, and are proportionately
thinner. The head is protected by dermal ossifications, the surface of which is covered
with enamel raised into, and arranged in, an exquisite series of parallel but sinuous
ridges. The external covering of the head is preserved in many specimens, but in all cases
more or less crushed and disturbed. The cranial bones include frontals (/), parietals (p),
and occipitals (o) : a restored outline of their forms, one half the natural size, is
attempted on PL XLIX. fig. 2, from which it may be inferred that the several bones

SECOND SERIES. — ZOOLOGY, VOL. II. 64

428 ME. J. W. DAVIS ON A NEW SPECIES OE CCELACANTHES

constituting the upper surface of the cranium form a compact and closely attached group.
Anterior to the frontal hones there is one forming the snout, prohahly the ethmoid (eih),
which readies over and more or less envelops the anterior extremities of the rami of the
upper jaw. The orbit (orb) is moderately large ; its upper boundary is formed by the frontal
and parietal bones : between these and the maxilla the structure of the lateral surface of
the head is somewhat obscure. There was an important osseous plate extending from
the orbit to the opercula, apparently the temporal bone (t) ; and one or more smaller
ossicles filled up the interval between the maxilla, the anterior portion of the orbit, and
the bones of the cranium. The opercular bones are three in number ; the operculum
(op) was a large bony plate, subtriangular in outline, extending to and overlapping the
scapular bones of the pectoral arch. In addition to the operculum there are two other
plates, similar in form, but only about one fourth its size : they were probably the sub-
operculum (i.op) and interoperculum (s.op), and are represented on PI. XLVII. fig. 1, and
also in the l'estoration represented on PL XLIX.

The upper jaws (mx) are frequently observed in specimens, but in no instance with
sufficient certainty and clearness to enable an exact description to be given of their form.
They were considerably shorter than the mandibles (M) of the lower jaw, thickest and
strongest beneath the orbit ; and, tapering towards the symphysis, the most anterior portion
was probably divided by the insertion of the preinaxillaries (PI. XLVII. fig. 11). The
mandibles are strong, about 3 inches in length ; the lower margin is straight, and was
grooved and smooth where attached to the jugular bones ; the upper margin is anteriorly
depressed, but expands towards the median portion to a height of '35 of an inch above the
inferior margin ; towards the posterior extremity the surface is again depressed. The pre-
sence of teeth in either the maxilla or mandible is somewhat problematical. An exami-
nation of many specimens has failed to show an example with teeth in situ. At the same
time, a number of teeth are occasionally seen mixed with the bones of the head, and
it is possible may have been so loosely attached to the jaws that they have invariably
become separated after the death of the fish. The teeth arc about "1 of an inch in
length, smooth, slightly curved, and acutely pointed. The whole of the under surface
of the head between the two rami of the lower jaws is occupied by the jugular plates (ju):
they are large, elliptical in form, internally strengthened by a median process traversing
the longitudinal axis of the plate, and externally ornamented by an extremely beautiful
arrangement of ridges on the surface of the ganoine, varying much in detail, but running
for the most part parallel with its longer axis.

The scales are moderate in size and rather thin. About one third of the surface of
each scale is exposed, the remaining portion being covered by the succeeding scales.
The outline of the exposed part forms a parallelogram, the exposed margin forming a
more or less acutely pointed angle, whilst the margin hidden by overlapping scales is
circular. The scales covering the ventral surface are the largest and broadest : they
measure -3 of an inch across; the length of the scale is -35 of an inch, of which the
exposed surface occupies '2 of an inch. The scales on the sides average the same length
as the ventral ones, but are only about 2 of au inch across. The scales decrease in size
towards the tail : they do not appear to exhibit any well-defined lateral line ; they run

FROM THE YORKSHIRE CANNEL COAL. 429

in parallel rows from the dorsal to the ventral surface, diagonally, with a slight sigmoidal
curvature. The scales enveloping the lohes of the several fins differ from those of the
body only in size ; they are much smaller, decreasing as they approach the distal ex-
tremity of the lobe.

The external surface is covered with glistening enamel, which is minutely furrowed.
The intervening ridges vary considerably in detail, both as to arrangement and form : in
some instances the ridges extend parallel with each side of the scale and meet at the
point, the successive ridges forming a series of triangles whose apices extend in a line
down the middle of the scale ; this is especially the case along the lateral surfaces of the
body. On other scales, especially along the ventral surface, the ridges extend more or
less along the axis of the scale, the longest being iu the middle, and becoming gradually
shorter on each side, and running out on the margins of the scale, giving it a some-
what jagged and rough outline. The ridges are not unfrequently broken by furrows
extending across them, and where this happens the surface has the appearance of
being covered with elongated tubercles. Various modifications of the several forms
described occur on the scales of a single specimen, and sometimes in close proximity one
to another (PL XLIX. fig. 3).

The fins are supported in each instance, with the exception of the anterior dorsal fin,
by a pedunculate, lobate expansion of the teguments of the body. The lobes of the
posterior dorsal and the pair of ventral fins are large : in the specimen depicted on
PI. XLVI. they are quite half an inch in diameter ; those of the anal fin and the pectorals
are not so large. The anterior dorsal fin exhibits the ordinary arrangements of the fin-
rays, though they appear to be attached, in the specimen already referred to, without the
intervention of a lobe, apparently in close connexion with the neural spines. It is
impossible to distinguish the exact number of rays which compose each fin ; but twenty
rays may be counted fringing the extremity of the lobe of cither the dorsal or the ventral
fins, and it is probable that this number is fewer than actually exist, because the rays
are more or less folded and bent under each other. The pectoral and anal fins were
composed of a smaller number. The rays vary considerably in length : those springing
from the extremity of the lobe are an inch and a quarter to an inch and a half in length,
whilst those above and below gradually diminish, the sbortest being less than half an
inch in length. The basal portion of each fin-ray, for about two thirds the entire length,
is formed of a single bone pierced by a hollow tube, like the spinous bones of the body;
the remaining third is divided by numerous transverse jointings into small ossicles.

The caudal fin is very large and powerfully built. The vertebral column extends in a
straight line from the body of the fish through, and some distance beyond, the caudal fin.
Its length beyond the extremity of the fin-rays shows a considerable amount of variation
in different specimens. In the one figured on PI. XLVIII. fig. 2 the vertebral prolongation
barely extends beyond the fin ; but in others, and in many instances smaller specimens,
there is fully an inch between the termination of the caudal-fin rays and the beginning
or basal portion of the smaller rays which form the second caudal appendage. Through-
out the whole length of the vertebral support for the second caudal fin there are small
rays on both the upper and lower surfaces, which appear to have served as a fin-like

64*

430 MR. J. "W. DAVIS ON A NEW SPECIES OP CCELACANTHUS

connexion "between tbe first and second caudals. The rays of the larger caudal fin are
similar to those already described in connexion with the fins of the body ; they are
formed of a single pierced ray, connected with the interspiiious bones to be mentioned
further on, the distal extremity of which is divided by a number of joints with articulating
surfaces. The rays are thirty-eight or more in number, half of which are above and half
below the median extension of the vertebral column. The second caudal also consists of
a numher of short rays, which form a duplicate in miniature of the larger caudal fin,
without the continuation of the vertebral axis beyond its extremity.

The specimen of which a figure is given on PI. XLVII. fig. 1 affords an almost perfect
representation of the structure of the under surface of the head. The anterior external
boundary of the specimen is formed by the right and left ramus of the lower jaw (m), each
3-3 inches in length. The space between the rami of the jaw was wholly occupied, when
the fish was living, by a pair of jugular plates (ju) ; these extended from a position imme-
diately under the point of contact between the mandibles backwards a distance of nearly
3 inches. The external lateral margins of the jugulars fit to and occupy a groove along
the under surface of the mandibles, whilst mesially one appears to have slightly
overlapped the other : this is indicated by there being no ornamentation on the inner
surface of the left jugular, whilst that of the right one extends to the margin of the plate.
On either side may be seen the operculum (op), a large subtriaugular bone 1*3 inch in
diameter. The under surface is represented in this figure ; the upper or external surface
is covered with a beautiful arrangement of stria?, very minute but distinct, feebly repre-
sented by a separate operculum on PI. XLVII. fig. 7. In addition to the opercula there
are two similar but smaller bones on each side, which were probably connected with the
operculum, and may have been suboperculum and interoperculum (s.op and Lop). For-
tunately the external hones have been pressed aside and separated, so that the internal
arrangement of the branchial arches (br) is exhibited. There are five branchial ossifi-
cations, extending diagonally from each side of a median support : they are strong, with
a more or less sigmoidal curvature, and vary in length between 1*7 and T2 inch, the
anterior arch "being longest, and those placed posteriorly gradually decreasing in length ;
they have a diameter of about "2 of an inch. The median support is a strong bone 2 inches
in length, "2 of an inch in breadth near the middle, widening out anteriorly to 3 of an inch,
and terminating in an acute point : posteriorly the surface expands to -5 of an inch, and
bifurcates at the posterior extremity. A detached specimen is represented by fig. 10 on
PL XLVII. Between the two jugular plates aud in front of the median branchial support
there is an ossified mass, which may represent the cruciform bone which Prof. Huxley
describes as forming the anterior extremity of the strong median ossification in Coelacan-
thus Upturns (Memoirs of Geol. Survey, decade xii. p. 19). In this instance, if the mass
represents the cruciform extremity, it is a separate bone. The opposite extremity also
differs from the enlarged figure in the Decades, which is described as a "posterior elon-
gated spatulate portion."

The specimen drawn on PI. XLVIII. exhibits a further extension of that portion of the
body forming the junction of the head with the trunk of the fish. The jaws and jugular
plates (ju) are exposed, and extending from beneath these are a pair of bones (x), one on

FEOM THE YORKSHIRE CANNEL COAL. 431

each side, which constitute a part of the pectoral arch and support the pectoral fins (p). The
bones are nearly 2 inches in length : their anterior extremities extend towards the median
portion of the branchiostegal apparatus, and are separated about an inch from each other ;
extending in a semicircle, the posterior extremities, somewhat flattened, are attached to
the pectoral fins. A process on the outer margin of the bones gave attachment to other
members of the pectoral arch, some elements of which are probably represented by the
specimens on PI. XLVII. figs. 4, 5, and 8. A separate example of the bone described
above is represented by fig. 12 on the same Plate.

The vertebral column is in no instance preserved, but a decided impression remains in
specimens showing its course through the body. The absence of bony vertebrae indicates
that they were entirely cartilaginous, and in consequence have not been preserved.
Neural spines extended from the vertebral column towards the dorsal aspect of the fish.
They are strong and closely set ; there are twelve in the space of one inch in that portion
of the body surmounted by the anterior dorsal fin ; they are 1 inch to lh inch in length,
being longest and most obliquely placed near the caudal extremity of the fish. The base
of the spine attached to the vertebral column was divided, so as to afford a passage through
which the spinal cord passed ; the bifurcation extends -3 of an inch. The rays are
hollow, the enclosed tube occupies about two fifths of the diameter of the spine, and the
encircling bone is of a very fine close texture.

Prof. Huxley, in the Memoirs of the Geological Survey, decade xii. pp. 1G and 18,
considers that the dorsal fins of 0. lepturus are supported by large interspinous bones.
Numerous specimens of the species now described have afforded no definite indication of
such an interspinous bone ; but rather seem to show a more close connexion between
the spinous processes and the rays of the fin by smaller interspinous hones.

The whole of the ventral portion of the fish, from the insertion of the pectoral fius to
midway between the ventral and anal fins, is occupied by an ossified air-bladder, and it
is worthy of remark that the whole of this region appears devoid of ribs or haemal spines :
bones are occasionally found lying across this region, but they appear to have been
displaced and fallen there accidentally. Posteriorly, haemal processes, corresponding to
those on the neural aspect of the vertebras, extend to and support the anal fin and the
inferior lobes of the caudal fius. The spinous processes are longer in this region than
elsewhere, and appear to be directly connected with the rays of the tail without the inter-
vention of interspinous bones (PI. XLVIII. fig. 2). The connexion between the neural
or h?emal spines and the fin-rays is not very clearly defined ; but, so far as can be
ascertained, there is no special arrangement such as was pointed out by the late Prof.
Agassiz (' Poissons Poss.' vol. ii. p. 168). He describes the interspinous bones as ossicles
which, instead of being attached by ligaments to the spinous bones on each side, are
attached to the end of one of them, so as to form its direct continuation ; and the ray,
properly so called, is stated to be forked at its base, and embraces the interspinous ossicle
in a manner similar to that of the spinous processes embracing the vertebrae. The
specimen referred to above exhibits very clearly the attachment of the caudal rays to the
spinous (?) bones. The basal extremities of the rays in this species are bent a little to
one side, and thin out to a point. The distal extremity of each spinous bone, which is

432 ME. J. W. DAVIS ON A NEW SPECIES OF CCELACANTIIUS

also pointed, is in juxtaposition with the ray ; but there is no appearance of bifurca-
tion on the part of the latter. The terminal extremity of each ray of the caudal fin is
divided into numerous little ossicles by transverse joints, as in the case of those of the
other fins. The vertebral prolongation extends to a varied extent in different specimens
beyond the first caudal fin : in the specimen now figured it terminates with a slightly
rounded extremity, coequal with the length of the fin-rays ; it is enclosed in enamelled
scales quite to the end, from which a number of small rays expand to form the second
caudal appendage. The number of rays is not easy to determine, but there are at least
twelve of them. In a second specimen, not so large, and apparently pertaining to a
younger fish, the vertebral continuation extends an inch beyond the termination of the
major fin ; its upper and lower surfaces are provided with short fin-rays, and the second
caudal fin extends from its extremity, as in the previously described example.

A portion of the suspensory apparatus is represented by the hyomandibular bone on
PL XLVII. fig. 3. There are also other bones whose position and function cannot be
satisfactorily ascertained.

The air-bladder was large, and extended from a position immediately behind the pec-
toral arch a distance of 7 inches, to a point immediately under the second dorsal fin. Its
depth is about 1 inch. It presents the appearance of a dull, semi-lustrous body, breaking
with an irregular fracture, and capable of division into a large number of very thin layers.
The latter are semi-transparent, something like thin layers of shellac. Under the micro-
scope, no very decided characters can be detected ; the layers do not present the appear-
ance of osseous structure, but perhaps resemble as much as any thing the chitinous layers
of the dermal covering of some insects. The outward appearance is well represented by
the artist on PL XLVL, extending along the ventral portion of the specimen over two
thirds its length, and on PL XLVIII. fig. 1 by an oval mass between the pectoral fins,
where the scaly covering of the fish has been removed. The structural divisibility into a
number of layers appears to indicate that the air-bladder consisted of an outer chamber
or sac, divided into a considerable number of compartments by longitudinal septa, these
also being, in all probability, subdivided by transverse partitions, which, however, have
not been distinguished in the fossil state.

The species of Coelacanthus now described is from the Cannel Coal in the Middle Coal
Measures at Tingley, in the West Hiding of Yorkshire, and with it have been found a
number of other fish-remains. In a paper read before the Geological Society, and published
in the Quarterly Journal of the Society for Pebruary 1880 *, a detailed description of the
stratigraphical position of the Cannel Coal, as well as the manner of its occurrence, is
given. In several respects this species possesses distinctive features, which exclude it
from either of those described by Agassiz (' Poissons Possiles,' vol. ii. p. 170), or from
those which arc described, or whose descriptions have been amplified, in the compre-
hensive treatise by Prof. Huxley in the Memoirs of the Geological Survey (decade xii.).
The most important characters exhibited by the species now described, which I propose

* Quart. Journ. Geol. Soc. vol. xxxvi. p. 50.

FROM THE YORKSHIRE CANNEL COAL. 433

shall be designated Coelacanthus Tingleyensis, consist in the method of attachment of
the rays of the dorsal and caudal fins to the spinous processes of the vertebrae, in the
larger number of the fin-rays, and in their extremities being articulated ; the orna-
mentation of the scales and exposed bones of the head consists partly of ridges and also
of minute tubercles ; and they also differ in the form of the scales.

DESCRIPTION OF THE PLATES.
Plate XLVI.

Coelacanthus Tingleyensis, Davis (nat. size).

Plate XLVII.

Fig. 1 . Coelacanthus Tingleyensis, anterior portion of the body, exhibiting the under surface of the head

(nat. size).
Figs. 2-13. Detached bones of the head.

Plate XLVIII.

Fig. 1. Coelacanthus Tingleyensis. Thoracic arch with attachment of the pectoral fins (nat. size).
Fig. 2. Caudal fin of same (nat. size) .

Plate XLIX.

Fig. 1. Restoration of C. Tingleyensis (half nat. size).
Fig. 2. Upper surface of head, restored.
Fig. 3. Scales, magnified.

[ 435 ]

XVII. On three new Species of Metacrinus. By P. Herbert Carpenter, D.Sc,
Assistant-Master at Eton College. With a Note on a new Myzostoma, by Prof.
L. von Graff, Ph.D. (Communicated by Br. W. B. Carpenter, F.B.S., F.L.S.)

(Plates L.-LII.)

Read 5th June, 1884.

TlIE dredgings of the 'Challenger' at four stations in the Western Pacific and the
Malay Archipelago yielded ten species of a remarkable new Crinoid allied to Penta-
crinus. Most of these were drawn for the Export on the Stalked Crinoids, under the
superintendence of the late Sir Wyville Thomson ; and when the collection came into
my hands after his death, I found the name Metacrinus in his handwriting upon a proof
copy of one of the plates. This name is consequently employed throughout the ' Chal-
lenger ' Reports, in which the peculiarities of the genus are fully described. I have else-
where noticed the one which is most obvious at first sight as specially distinguishing it
from Pentaerinus*, viz. the presence of four, or even of sis radials, instead of three only.

The other special marks of the genus are : — 1. The large size and somewhat cuboidal
form of the basal joints of the lower pinnules ; 2. The derivation of the ambulacra of the
lowest pinnules, borne on the outer radials, either directly from the peristome or from
the five primary ambulacra proceeding from it ; 3. The modification of the supranodal
stem-joints, as well as in most cases of the infranodal joints. All these characters are well
shown in the three species described in the following pages.

The fine species for which the name Metacrinus rotundus is proposed (PI. L. fig. 1)
was dredged at a depth of 70 fathoms in Sagami Bay, Japan, by Br. L. Boderlein, the
Conservator of the Natural History Museum at Strassburg, in the year 1881 1- He
has been kind enough to intrust it to me for description, together with a considerable
variety of Comatuhe (both Antedon and Actinometra), which he found to be extremely
plentiful in the same locality. These will be considered later; but I would here record
my indebtedness to Br. Boderlein, and beg him to accept my cordial thanks for his
kindness.

A fine Myzostoma, which I found attached to the Metacrinus, is figured and described
by my friend Prof. L. von Graff, of Graz. I could find none of these parasites upon the
Comatulce, nor were there any loose in the spirit in which they were preserved.

The two other species of Metacrinus have been for some years in the collection of my
friend Mr. Charles Stewart, F.L.S. He obtained them from the officers of the Eastern
Telegraph Company at Singapore, where they were brought up in the process of picking
up a cable for repair. They were sent home in spirit ; but, by some unfortunate blunder
on the part of an ignorant clerk, were taken out and dried upon their arrival in this
country. The natural result was, in the case of Metacrinus superbus, that a considerable

» Bull. Mus. Comp. Zool. vol. x. 1882, p. 1G7.

t " Fauuistiseho Studion in Japan, Enoshima und die Sagami-Bai," Areuiv fur Naturgesch. 49 Jahrg. p. 119.
SECOND SERIES. — ZOOLOGY, VOL. II, 65

436 DR. P. II. CARPENTER ON

number of the arms broke at one of the lower syzygies, and the great beauty of a large
and perfect Crinoid was irretrievably ruined.

The third species herein described, Metacrinus Stewarti, is unfortunately only repre-
sented by a stem fragment ; but its characters are so different from those of the stem in
any species of Metacrinus yet known to me, that I have no hesitation in regarding it as
belonging to a new specific type.

Although Mr. Stewart had long since commenced to write an account of the fine
specimen in his hands, and had also made a few drawings of it, his usual unselfish kind-
ness has led him more than once to offer it to me for description ; and when I received
Dr. Doderlein's specimen, it seemed to me to be a good opportunity to avail myself of
Mr. Stewart's generous offer, and to describe the two species together.

1. Metacrinus rotundus, sp. nov. (PI. L. ; PI. LII. figs. 1-7.)

Dimensions.

Length of stem to twenty-ninth node 44 centimetres.

Diameter „ 5 millimetres.

Longest cirrus, 45 joints \7 „

Diameter of disk IS ,,

Length of arm after palmar axillary, 130 joints 125 ,,

„ radial pinnule, 20 joints 20 „

„ distichal pinnule, 20 joints 20 ,,

„ first pinnule after palmar axillary, 1 1- joints ... 9 „

Stem robust, with a rounded pentagonal outline. Usually eleven or twelve, but some-
times as many as eighteen intcrnodal joints. They have fairly well crenulated edges,
and their sides bear faint, more or less interrupted ridges, which sometimes become
slightly tubercular at the angles. These interradial tubercles are more marked on the
nodal joints, but vary considerably in size. The nodal joints increase in size from above
downwards as far as the lower edges of the wide cirrus-sockets, and then diminish again.
The sockets extend upwards above the articular surface on to the supranodal joints,
which are thus somewhat incised ; but each socket terminates below by a well defined
lip, which is distinctly above the lower edge of the nodal joint. The infranodals are not
incised to receive the cirrus-bases, so that the apposed syzygial surfaces are almost per-
fectly circular. The cirri consist of 40-45 very uniform squarish joints, the lowest of
which are but little wider than their successors. The cirri at the twelfth and next
following nodes are larger than those lower down. The intcrarticular pores disappear
between the eleventh and twelfth nodes.

Basals pentagonal, rounded, and prominent. Padials usually five, with a syzygy in
the second, and occasionally another in the fourth or axillary.

The rays divide three, or sometimes four times, giving 40-50 arms, which consist of
about 130 joints beyond the palmar axillary. They are tolerably smooth and but little
serrate in the mediodorsal line, except near the ends.

Primary arms of 7-10, usually eight, distichal joints ; secondary arms of 10-18, usually
twelve or fourteen, palmar joints. There is occasionally another axillary after twenty or

THKEE NEW SPECIES OF METACPJNUS. 437

twenty -two joints more. There is usually a syzygy iu the third joint after each axillary.
The next is somewhere hetween the twelfth and thirtieth brachial, and others follow at
very irregular intervals.

The three radial pinnules, on the second, third, and fourth radials, are large and pro-
minent, reaching 20 millims. in length, and consisting of 18-20 joints. The three lower
ones are massive and cuboidal ; the next few flattened laterally, but still stout, and pro-
jecting beyond the bases of their successors. This is most marked in the smaller terminal
joints, so that the end of the pinnule lias a very serrate appearance. The first distichal
pinnule is much like those on the radials ; but the following ones are smaller and less
serrate. The two lower joints are flatteued, and much wider than their successors ; but
this inequality is much less marked after the palmar axillary, and disappears altogether
in the later pinnules, which are comparatively small.

The disk is covered by an irregular pavement of small plates, set moderately close,
and the ambulacra are well protected by plates; but the perisoine at the sides of the
disk, between the rays, is almost bare.

Neither are the muscular bundles between the arm-joints covered by plates, which are
limited to the ambulacra. They form a fairly regular row on each side, and become
differentiated at the bases of the pinnules into more or less pointed side-plates and
rounded covering-plates. Colour, a very light yellowish brown, which has scarcely
altered in spirit.

Hub. Sagami Bay, Japan ; 70 fathoms.

Remarks. — Several well defined characters distinguish this fine species from the various
types of Metucrinus which were dredged by the ' Challenger.'

In the first place, like M. Moseleyi, it occupies an intermediate position between the
two groups into which most species of the genus very naturally fall : — 1. Those with
four radials, of which the second is a syzygy ; and 2. Those with six radials, of which
both the second and fourth are syzygies. Variations from these numbers occasionally
present themselves in single rays of individual specimens ; but, as a general rule, the
presence of four or six joints in a ray is a very constant character.

This is not the case, however, in 31. Moseleyi, two individuals of which were dredged
by the ' Challenger.' The following variations occur among their ten rays : —

1 ray of three joints, with the second and the axillary a syzygy.
i ,, ioiu ,, ,, ,, ,, ,, ,,

3 rays of five ,, ,, ,, a syzygy.

2 ,, ,, „ ,, ,, and the fourth a syzygy.

1 ray of six ,, ,, ,, a syzygy.

2 rays „ „ n ;) and the fourth a syzygy.

In the single specimen of M. rotundas the construction of the five rays is as follows : —

1 ray of six joints, without a syzygy.

2 rays of five joints, with the second a syzygy.

1 ray of five joints, with syzygies in the second and fourth.

1 ray of five joints, with syzygies in the second and axillary. r

G5

r.*

438 DR. P. H. CARPENTER ON

In three of the rays, therefore, there were six primitive joints, of which the second
and third became ultimately united by syzygy in two cases, but remained separate in a
third ; while in the other two rays there were primitively seven joints.

The type, therefore, is intermediate between those more regular species, such as M.
angulatus from the Arafura Sea, with five primitive radials, two of which eventually
unite by syzygy, and the other group typified by 11. interruptus from among the Philip-
pine Islands, which have eight primitive radials that form two syzygial pairs.

It is to the species last mentioned that M. rotundus presents, on the whole, the most
resemblance. There is about the same number of joints in the primary and secondary
arms, and the external characters of the stem are very similar in the two types. The
number of internodal joints is almost the same ; but the horizontal ridges on their surface
are much less marked in M. rotundus than in M. interruptus, while their outline is more
rounded and less sharply pentagonal (PI. LII. figs. 7, 5).

But it is in the characters of the nodal and intranodal joints that M. rotundas differs
from M. interruptus, and still more so from all other species of the genus.

In most Pentacrinidse the cirrus-facet is Limited to the nodal joint, but lies at the
broad end of a pear-shaped socket, the narrower part of which is continued downwards
on to the intranodal joint, and receives the base of the cirrus. The upper part of this
socket in Pentacrinus is confined to the nodal joint, not extending above the edge of the
cirrus-facet, which is often considerably lower than the vertical height of the nodal
joint; while the cirri are mostly directed downwards, and the infranodals are more or
less grooved to receive their bases, as described above. In Metacrinus, however, the
normal direction of the cirri seems to be upward rather than downward (Pis. L., LI., LII.
fig. 13) ; and the cirrus-sockets, therefore, encroach somewhat upon the supranodal,
which is thus slightly incised, as may be seen by comparing figs. 1 and 1 on PL LII. The
former represents a normal internodal joint, and the latter that immediately above the
node. The share of the supranodal joint in forming the cirrus-socket is likewise seen in
the side-view of the stem (PI. LII. figs. 5, 12, 18). In most species of Metacrinus the
intranodal is also slightly grooved to receive the cirrus-bases, and its upper (syzygial)
surface is, therefore, more or less stellate, in correspondence with that of the nodal joint
above it. This is least marked in M. interruptus, the nodal joints of which differ but little
in outline from those in other parts of the stem. The cirrus-facet ends below in a well
defined rim, which is distinctly above the lower edge of the joint. The syzygial surface,
therefore, has almost exactly the same pentagonal form as the ordinary sculptured face
of an internodal joint. That of the intranodal is similar to it, and has no reentering
angles as is usually the case.

Now in M. rotundus there is the same well-marked termination of the cirrus-facet
above the lower edge of the nodal joint (PI. LII. fig. 5), and neither it nor the infra-
nodal is in any way grooved to receive the cirrus-bases (PL LII. fig. 3). Their apposed
surfaces, however, are not pentagonal, as in M. interruptus, but they are almost circular
(PL LII. fig. 3); and this character distinguishes the stem of M. rotundus from that of
all the other species of the genus.

The type which approaches it most nearly in this respect is also a Japanese form, viz.

THREE NEW SPECIES OF METACBINUS. 439

that dredged by the ' Vega ' in 1879, at a depth of Go fathoms, in the Bay of Yeddo.
Prof. Loven has been kind enough to send me some fragments of the stem, and also to
permit my friend Mr. "W. Percy Sladen, P.L.S., to make an investigation of the type
during his recent visit to Stockholm for the purpose of examining the collection of Star-
fishes in the Museum. I am greatly indebted to Mr. Sladen for the trouble which he
took, and for the careful and detailed description of the ' Vega ' specimen with which he
has furnished me. It has nearly the same number of internodal stem-joints as M. inter-
rupt us and M. rotundus, and the length of the primary and secondary arms is very much
the same as in these types. In the number of radials (six, with syzygics in the second
and fourth) it agrees with M. interruptus and differs from M. rotundus ; but it approaches
the latter form more nearly in the characters of the stem.

Although the diameter of the stem-joints is less than in M. Interruptus, and still more
so than in J/, rotundus, they are not only relatively but absolutely higher than in these
species. They resemble the latter rather than the former in the slight amount of sculp-
ture on their sides, and in their more rounded pentagonal form. As in both these types,
the cirrus-sockets end well above the lower edge of the nodal joint ; but the syzygial
surface, though much less sharply pentagonal than in M. interruptus, is far from being
as regularly circular as in M. rotundus (PI. LII. fig. 3).

In the robustness and other characters of its stem, therefore, M. rotundus is distin-
guished from the two species which seem to resemble it most, just as it is by the irregu-
larity in the number of radials. The plates of the cup and arms are altogether larger
than in most species of the genus, the only ones which approach it in size being AT.
angulatus, M. cingulatus, and M. nobilis, all from near the Ke Islands (Station 192).

The pinnules borne by the radials and lowest distichals have especially stout joints ;
and the serrate appearance of their outer extremities, produced by the elevation of the
distal edge of each joint, is extremely well marked, so as to recall the distinctive features
of Pentacrinus asterius.

The disk of M. rotundas is excellently preserved, and shows very well one of the special
characters of the genus, viz. the origin of the ambulacra of the large lower pinnules either
directly from the peristome, or from the five primary groove-trunks before they bifurcate
(PI. L. fig. 2). The intcrpalmar areas of the disk are paved with small plates, which are
pierced by the water-pores, but arc not quite so closely set as they arc in some of the
Pentacrinidoe and Comatuke; while there are few or no plates in the perisome at the
sides of the disk between the rays. The disk-ambulacra, however, are protected by
several irregular rows of plates, which in some places meet so closely over the middle
line of the groove that it is converted into a tunnel. This was the case, not only
on the arms, but also in the calyx-ambulacra of the Pakeocrinoids, whether the latter
were external, as in Cyalhocrinus, or still further covered by a solid vault, as in Actino-
crlnus.

The arms have no anambulacral plates overlying the muscular bundles, as is so often
the case ; but the ambulacra are well protected by a double row of somewhat irregular
plates, in which there is more or less distinct evidence of bifurcation. This becomes
specially marked at the bases of the pinnules, where the proximal limb of the fork

440 DR. P. II. CARPENTER ON

remains as the side-plate, with a squarish hase and pointed top ; while the distal limb of
the fork becomes the rounded covering-plate (PI. L1I. figs. 6, 7). As in all the Penta-
crinidse, the two series end some little way before the extremity of the pinnule, the
ventral groove of which is thus left entirely unprotected. It is rare to find any indi-
viduals of Metacrinus some part of which does not exhibit irregularities of growth ; and
M . rottmdus is no exception to the rule, the seat of the irregularity in this case being
the basal ring. One of the basals is much wider than its fellows, and looks as if it were
divided into two unequal parts, one of which slightly overlaps the other. This may be
i\uti either to an anomalous mode of growth from the beginning, or else to fracture and
subsequent reparation. The point is not one of very great importance, but is only of
interest from the frequent occurrence of similar irregularities in Metacrinus and their
comparative absence in Pentacrinus, which seems, as it were, to be somewhat more crys-
tallized in the regularity of its characters.

The fine specimen of Myzostoma cirripedmrn (PI. LII. fig. 19), which is described
further on by Prof, von Graff, was attached to this individual; while its cirri afforded
anchorage to those of a small Antedon.

Dr. Doderlein informs me that it was attached to the tangles, and had the arms closed
in over the disk when it was first removed from the water, but that they opened out
after a time. Most of them eventually broke off at the syzygy in the third distichal
after the specimen bad been put into spirit. Its colour, however, has altered but little
in consequence.

2. Metacrinus mperbus, sp. nov. (PI. LI. ; PL LII. figs. 8-12.)

Dimensions.

Length of stem to twenty-sixth node 30 centimetres.

Diameter ?■?.-, millimetres.

Longest cirrus, 68 joints 80-00

Diameter of disk 30-01)

Length of broken arm after suprapalmar axillary, 100 joints . 105*00 „

Length of first palmar pinnule, 27 joints 36"00 ,,

„ „ pinnule after suprapalmar axillary, 22 joints . 18'00 „

Stem very robust, with a rounded pentangular outline and smooth flat sides. 9-11
internodal joints, with moderately crenulated edges. The angles of the nodal joint*
scarcely project at all, and the wide cirrus-sockets between them extend well up on to
the supranodals, though but very slightly on to the infranodals. The cirri increase in
size down to the seventeenth node, and then remain tolerably uniform. They consist of
about 05 stout, but short and wide joints. The intcrarticular pores disappear at the
eighteenth node.

Basals prominent, and widely pentagonal, with downward extensions over the upper
stem-joints. Radials wide, usually four in number, with a syzygy in the second.

The rays divide four times, and the outermost of each pair of the arms which are
borne by the suprapalmar axillary divides again, so that the number of arms must exceed

THREE NEW SPECIES OF METACRINUS. 441

100. There are over 100 joints above the suprapalmar axillary, all of them with raised
distal edges. This character is not specially prominent on the later joints, but is very
well marked on the outermost radials and on the lowest arm divisions.

Primary arms usually of 4-6, but in one case of 10 distichal joints. 7-10 palmars in
the secondary arms. Tertiary arms usually of 14-16 (rarely 12 or 18) joints, with
another axillary after 16 or IS (rarely 12, 20, or 22) joints on the outermost of each pair
of quaternary arms. The third joint after each axillary is a syzygy ; and the second
syzygy in the free arms may be anywhere between the thirtieth and sixtieth brachial.
Others follow at intervals of 8-20 joints.

The radial, distichal, and palmar pinnules are all large and styliform, especially the
two latter. That on the second radial has two large and cuboidal basal joints ; but the
following ones become rapidly smaller and much flatter.

The first distichal pinnules are much longer again, and consist of nearly 30 joints, of
which the basal ones, including even the first two, are much flattened laterally, though
of considerable depth. The following ones are of about the same length, but have wider
and more massive basal joints. After the first palmar pinnule the size gradually dimi-
nishes, rapidly at first, and then more slowly, while the pinnules become more flattened,
though the enlargement of the basal joints is visible for some distance out into the arms.

The disk is paved by small irregular plates, which are not very closely set; but the
ambulacra are well protected by plates. The brachial ambulacra are limited to the
centre of the arm-furrow, and more or less differentiated on the pinnules into side- and
covering-plates. Colour in the dry state, a light purplish grey, with a greenish tinge on
the tips of the pinnules.

Sab. Singapore.

This magnificent specimen is the largest Metacrinm. and, in fact, the largest recent
Pentacrinite that I have yet seen. Few species have a stem exceeding 5 millims. in
diameter. It reaches 7 millims. in Pentacrinus asterius and Metacrinm nobilis; but the
stem of M. superbus is considerably wider than that of either of these two types. The
former is the only one which at all approaches it in the number and stoutness of the
cirrus-joints, as also in the frequency of the ray-divisions.

The number aud very regular grouping of the free arms in Metacrinm superbus is a
somewhat striking character. Palmars appear to be universally present, so that there
are 40 tertiary arms ; and, with one exception, all of them which are preserved, either
wholly or partially, divide in the same way. Each of them has a suprapalmar axillary ;
and of the two arms which this bears, the outer, except in one case, divides again ; so
that each palmar axillary bears 6 arms : 2, 1, 1, 2. If this arrangement extended all
round the disk, there would have been 4 x 6 or 21 arms on each ray, making a total
of 120.

It is exactly the same arrangement as often occurs on the distichal axillary of Penta-
crinus asterius, which has 12 arms to the ray; and also on the radial axillary of P.
Mulleri, P. Maclearanus, and P. Wyville-Thomsoni, It is of some interest, as affording
a clue to the mode of development of the armlets of the Jurassic Extrv.crinus, as 1 have
explained in the ' Challenger ' Report.

4.42 DR. P. H. CARPENTER ON

Metacrinus superbus is somewhat closely allied Loth to 31. Murrayi and to 31. nobilis,
having, like them, a smooth flat stem and normally only four radials, and also ahout the
same number of joints in the primary and secondary arms. It has slightly shorter inter-
nodes than both these types ; and the joints composing them arc less sharply pentagonal
than in 31. nobilis, though the sides are not quite so much incurved as in 31. Murrayi.
Both the supra- and the intranodal stem-joints are about as mucli incised as in the latter
species, and somewhat less so than in 31. nobilis. In neither of these species are there
more than about 50 cirrus-joints, and the interarticular pores end at about the tenth or
twelfth node; whereas in 31. superbus there are some 65 cirrus-joints, and the inter-
articular pores extend down to the eighteenth node — a most unusual distance from the
calyx.

Both these types, again, differ altogether front the larger 31. superbus in the smoothness
of the dorsal surface of the skeleton as far as the middle of the free arms; while the
terminal portions of their arms are distinctly more serrate than in the corresponding
parts of 31. superbus.

Two of the radial scries are slightly irregular in the specimen now under consideration.

The type consists of live primitive joints, the second and third of which have united
to form a syzygy ; so that I he ultimate number of radials is four, with the second a
syzygy. On two rays, however, there were six primitive joints. In the one case the
second and third united as in the type, while the rest remained free; so that there are
two joints between the syzygy and the axillary, as shown in PI. LI. fig. 1. In the
next ray, however, the axillary became united by syzygy to the joint beneath it, and the
proximal and distal edges of the hypozygal are raised and thickened, as in the other parts
of the ray, but only for rather more than half their width. This is also the case with
the proximal edge of the axillary or epizygal, the smaller half of which appeal's as a
simple syzygial line, while the remainder is raised and thickened.

The disk, though mutilated in parts, is very well preserved for a dry specimen, and
presents one or two points worth notice. The peristome is large, but has no visible oral
opening, and a large number of food-grooves converge upon it, some of which proceed
direct from the large lower pinnules, as is always the case in this genus. There is, there-
fore, a great number of small interambulacral areas, which are paved by small plates
that are pierced by numerous water-pores, but not so closely set as to form a continuous
pavement. These perforated plates extend over on to the dorsal side between the ray-
divisions, just as they do in Pentacrinus asterius; though Miiller denied their existence in
this position *. The skeleton of the pinnule-ambulacra varies somewhat in appearance,
according to the size of the pinnule which bears it (PI. LIL figs. 10, 11). The side-plates
are usually low and rounded, and very different in appearance from the sharply pointed
plates which correspond to them in 31. rotuudus. The gradual development of the
pinnule-ambulacra from that of the arm is well shown in PL LI. fig. 2. In this, as
in all species of the genus, the ambulacra do not reach the distal ends of the pinnules,
but terminate some little way short of them ; and in like manner the terminal portions

* " Uobcr den Lau des Pentacrinus Caput-Medusm," Abhandl. d. Berlin. Akad. 184:?, p. 49.

THREE NEW SPECIES OF METACRINUS. 413

of the arms arc not only entirely devoid of any ambulacral skeleton, but they bear quite
small and rudimentary pinnules, which consist of only two or three minute joints, and
appear to remain permanently in this aborted condition (PI. LI. fig. 1).

Some of the arms, however, have been broken at a syzygy and subsequently repaired,
so as to be still in a state of growth, as is shown in PL LI. tig. 5. The manner in
which their terminal portions, and those of the growing pinnules, are coiled up, as it
were, is very singular, and forcibly recalls the characters of the arms and pinnules in
Holopus.

3. Metacrinus Stewarti, sp. now (PI. LII. figs. 13-18.)

Stem robust and pentagonal, the internodes consisting of 11 joints, with moderately
crenulated edges. The joints are G millims. in diameter, and have slight horizontal
ridges on their sides, which are continuous round the angles, so that these are slightly
produced. The nodal joints increase in size from above downwards as far as the lower
ednes of the wide cirrus-sockets, and then diminish again. The intranodal joints arc
not grooved at all to receive the cirrus-bases ; while the suprauodals are markedly
incised, and contribute to form the upper portions of the cirrus-sockets proper. The
cirri are 50 millims. long, and consist of 50 stout joints, the basal ones being much
wider than their successors, which are all uniformly cuboidal.

Remarks. — The characters of this stem-fragment are so well defined that I have no
hesitation in regarding it as belonging to a species of 3Ieiacriuus distinct from any yet
known ; and I am glad to have the opportunity of associating it with the name of my
friend Mr. Charles Stewart, P.L.S., to whom I am indebted for the opportunity of de-
scribing it.

In the length of the iuternodes and in the peculiarities of the nodal joints it resembles
the stems of Metacrinus rotundas, 31. interruptas, and the ' Vega ' specimen. It is larger,
however, than all of these, especially the two last, and the joints have much more distinct
horizontal ridges. These give it a certain amount of resemblance to the stems of 31.
Wyvilliia.rL& 31. cingulutus, though it differs altogether from these two in the characters
of the nodes.

Owing to the headless condition of the fragment, I was at first sight somewhat doubtful
as to which way up it should be placed. The line of union of the supranodal and the
nodal joints is rather less crenulated than usual, and looks almost like a syzygial line
between nodal and intranodal joints (PL LII. figs. 13, 18).

The nodal joints (PI. LII. figs. 14, 18) arc enlarged, and increase in size from both ends
towards the cirrus-socket, which extends on to what I now know to be the supranodal ;
but it seemed possible at first sight that this might really be the intranodal, and the stem
belong to a Pentacrinus with the cirri directed downwards, instead of a 3Ietacriuus with
upward turned cirri, as is really the case. In fact, the nodal joints have no little resem-
blance to those of Pentacrinus Wyville-Thomsoui, widening downwards to the lower edge
of the cirrus-sockets, and then falling away again (PI. LII. fig. 18). This is also
the case in Metacrinus rotundas, though to a somewhat less extent (PL LII. fig. 5), and
the upward extension of the sockets on to the supranodal is somewhat more marked than

SECOND SERIES. — ZOOLOGY, VOL. II. 66

444 PROF. L. VON GEAFF ON

in the Japanese species. The supranodal is not merely grooved to receive the cirrus-
bases, as in many species of the genus ; but it actually contributes to form the curved
upper edge of the socket itself, which is thus not entirely limited to the nodal joint.

Traces of this arrangement appear in Metacrinus rotundas (PI. LII. figs. 4, 5), though
they are far less marked than in M. Stewarti.

In consequence of the upward direction of the cirri, and the termination of the
cirrus-sockets well above the lower edges of the nodal joints, the infranodals are not
grooved at all to receive the cirrus-bases (PL LII. fig. 10) ; so that there are no re-
entering angles in their regular pentagonal outline, which is quite different from the
circular syzygial surface in Metacrinus rotundus (PL LII. fig. 3), and resembles that of
.1/. interruptus. The syzygial surface of the nodal joint, however, is somewhat lobate,
as in the ' Vega ' specimen. This lobate shape is partly the result of the increase in the
width of the nodal joint from above downwards, and its diminution again below the level
of the cirrus-socket by the downward slope of its outer surface, which thus becomes
visible in a view of the joint from beneath, as shown in PL LII. fig. 15. This is
especially marked in the lower side of the figure, and produces a corresponding incision
of the infranodal joint, which must not be mistaken for that produced by the downward
extension of the socket to receive the base of a cirrus. This, in fact, really does occur
in some abnormal sockets, as shown in the lower part of PL LII. fig. 16.

When all these peculiarities are considered together, it will lie seen that the stem of
Metacrinus Stewarti has a distinct character of its own. What its calyx may have been
is as yet unknown. That of M. interruptus, and also that of the ' Vega ' specimen, has
six radials, while M. rotundus has five (PL LII. fig. 1) ; and it is therefore uncertain
in which group M. Stewarti should be placed, while it might possibly even belong to the
group which includes M. superbus (PI. LI. fig. 1), with only four radials.

A large part of one side of the specimen, including many of the cirri, is wrrapped up
in a dense mass of what appears to be the remains of a horny sponge, in which are
imbedded Foraminifera, broken fragments of shells, and a quantity of earthy and mineral
matter. Some of the cirri afford an attachment to sessile Cirripcdes ( Verruca ?), as
shown on the left-hand side of PL LII. fis. 13.

Description of a new Sj)ecies of Myzostoma. By Prof. L. vox Graff, Ph.D.

Myzostoma cieripeditjm, sp. nov. (PL LII. fig. 19.)

A single wrell-preserved individual was found by Dr. P. H. Carpenter attached to the
specimen of Metacrinus rotundus which is described by him in the preceding pages.

Its external form resembles that of Myzostoma Wyoille-Thomsoni, mihi *, which infests
Metacrinus costatus, P. II. C, and M. angulalus, P. If. C, and has beeu described by
myself in the ' Challenger' Report.

* Zoology of the ' Challenger' Expedition, part xxvii. pp, 45-40, pi. vi. figs. 1, 2.

A NEW SPECIES OF MTZOSTOMA. 445

The very delicate body is in the form of an oval plate, 4-6 millims. long and 3 rnillims.
wide. Its thickness is very slight, hardly as much as in the European Myzostoma cirri-
ferum ; and it is therefore fairly transparent, not only in the clear yellow marginal
portion, but also in the somewhat thicker and darker central part within the ring of
parapodia. The edge of the body is so much folded over towards the ventral side that
the animal appears convex when viewed from above, and concave as seen from beneath.

The unusually extensive ramifications of the digestive canal * almost reach the edge
(PI. LII. fig. 19) ; and the marginal portion, which is free from them, is not sharply
marked off from the rest of the disk.

There are 20 cirri (<?.), disposed at tolerably regular intervals from one another, except
that the ninth and tenth cirri on each side, and also the two of the tenth pah*, are some-
what more widely separated. As regards the length of the cirri, the last pair, with a
length of 077 inillim., and the first pair, 0'6 millim., are the best developed. The lateral
cirri vary in length from (>14 millim. to 0 46 millim. The ventral longitudinal furrow
is well defined in most of them.

The body-wall bends slightly outwards between every two cirri, as is especially evident
in those parts which are not folded over; and the cirri, which are really marginal, tbus
come to lie in little bays within the edge.

The circlet of relatively weak parapodia (p.) is situated about halfway between the
centre and the edge of the disk, though the fifth pair is much farther from the hinder
end than the first pair from the front end. The parapodia reach 077 millim. in length,
when extended ; but there is nothing special either about their form or about the hooklets
which they enclose. On the other hand, there is a peculiarity of structure, hitherto
observed in no other Myzostoma, in the finger-shaped process (p.c.) which each para-
podium bears at its base on the side turned towards the middle line of the body. This
process is markedly distinguished from the marginal cirri by the want of a glutinous cell-
groove (" Klebzelle "). It is therefore best designated as a parapodial cirrus ; and its
function is probably that of a tactile organ, as with the similarly shaped structure in the
Chaetopods. The 8 suckers (s.) occupy a line situated about halfway between the bases
of the parapodia and the edge of the disk. They are very flat, and but slightly pro-
minent, having a somewhat elongated shape, and reaching 023 millim. in length and O'l
millim. in width.

Almost at the same level with the suckers is the ventral mouth-opening (m.), situated
0-35 millim. from the anterior end of the body. It leads into a considerably retracted
pharynx, the bulbus musculosus of which (ph.) reaches the unusual length of T23
millims. I could not make out the origin of the individual intestinal diverticula from
the central stomach (st.) ; but the rectum (»■.) is easily distinguished, together with the
oviduct (od.), which lies above it, and is filled with ova. The cloacal opening (cl.) is
placed terminally in a slight incision at the hinder end of the body. The two openings
of the male genital organs ( 6 ) are simple insignificant clefts close to the outer side of
the bases of the third pair of parapodia.

* These are indicated in a semidiagranrmatic manner on the left side of the figure oiuy.

446 CAEPENTEE AND VON GEAFF OX METACBINUS AND MYZOSTOMA.

DESCRIPTION OF THE PLATES.

Plate L.

Metacrinus rotundus, sp. n.
Fig. 1. The bead and upper part of the stem. Natural size.

Fig. 2. The disk, from above. x 3.

Plate LI.

Me/acri/nix superbus, sp. n.
Fig. 1. The head and upper part of the stem. Natural size.
Fig. 2. Portion of an arm, rather below its middle, from the right side. x 2.
Fig. 3. Terminal portion of an arm undergoing regeneration, showing the growing points of its two

divisions, x 2.
Fig. 4. Terminal portion of a full-grown arm, with imperfectly developed pinnules. x 2.
Fig. 5. The ambulacra! groove, with its branches on to the pinnules, viewed from above. x 5.

Plate LI I.

Figs. 1-7. Metacrinus rotundus, sp. n.
Figs. 1-4. Stem-joints, x 0. Fig. 1. Ordinary internodal joint. Fig. 2. Nodal joint ; lower or syzygial

face. Fig. 3. Nodal joint ; upper face. Fig. 4. Supranodal joint ; lower face.
Fig. 5. Fragment of stem, showing the cirrus-sockets on the nodal joiut. x 2.
Figs. 6 & 7. Side views of pinnule-ambulacra. x 25.

Figs. 8-12. Metacrinus superbus, sp. n.
Fig. 8. Nodal stem-joint ; upper face. x 4.
Fig. 9. Ordinary internodal joiut. x 4.
Figs. 10 & 11. Side views of piunule-ambulacra. x 25.
Fig. 12. Fragment of stem, showing the cirrus-sockets on the nodal joint, x 2.

Figs. 13-18. Metacrinus Stewarti, sp. u.
Fig. 13. The stem-fragmeut. Natural size.
Figs. 14-17. Stem-joints. x 6. F'ig. 14. Nodal joint; upper face. Fig. 15. Nodal joint; lower or

syzygial face. Fig. 16. Intranodal joint ; upper or syzygial face. Fig. 17. Ordinary internodal

joint.
Fig. 18. Fragment of stem, showing the cirrus-sockets on the nodal joint.

Fig. 19. Myzostoma cirripedium, sp. n., seen from the ventral side, x 28. t. Marginal cirri ; cl. Cloacal
opening ; *. Intestinal ramifications (indicated only on the left side of the figure) ; m. Mouth ;
od. Oviduct; p. Parapodia; pc. Parapodial cirri; ph. Bulbus musculosus of the Pharynx; s.
Suckers; st. Stomach ; $ . Male genital openings.

[ 447 J

XVIII. On the Breeding of Salmon from Parents which hare never descended to the Sen.

By Ebancis Day, F.L.S.

(Plates LIII. & LIT.)
Read March 5th, 1885.

CONSIDERABLE attention has been paid in this country to the life-history of the
Salmon, bnt many so-called "facts" heretofore accepted and that std pass current
have no better foundation on which to rest than the dogmata, assertion of a self-
led ..practical" individual, the theory of an ichthyologist or the unsupported
tenement of a museum naturalist ; or, .as Itussel • observed, - the nonsense about he
Salmon that has been published under the name of natural history, and thrust down the
Lets of Parliamentary Committees, is, when looked back upon, appalling » amount,

™tt~X«^ oi great moment that the various asserted facts upon
wUehany dXstuflingers 'should be accurately sifted, not by a partisan of one party or
"nether but an independent individual who would, regardless of expense tune and
trouble undertake suih a task in the sole interest of scientific truth and for he good of
he fi herfe Such tests have been for the last few years and are stdl being steadily ana
memitt nglv carried on at Howietoun, and in due time the results, whether m
"ec„Lnce\-ith or opposed to existing theories, will, it is to be hoped, be grven to the

"trteminent among the questions of practical moment, as bearing upon the race of
so-called "land-locked Salmon," is whether the Salma salar can be permanently returned
and finally breed in fresh water without descending to the sea.

I propose examining this principally from a fisb-eulturisfs point of view as, owing to
the Snl of the owner of Howietoun, I have had the "fT*^ ££Z f C
of the interesting .and instructive experiments which are being carried on n that large

d Cnlcted establishment, and consequently am in a position to give a rep y o

f QC 4f w there been solved. I have also thought that it would be

SereaTwI I'to c e ^the Opinions which have been published in the British

its on tnis subject, as well as a synopsis of a few of the various experiments which haie

"«" p'Shedthe first edition of his - Compleat Angler,' wherein
theopinl oftbemost reliable authors of previous or contemporary times on Salmon-
toeelgare condensed. He remarks that the Salmon - is said to breed or cast its spawn
• I "mstVivers in the month of A,,„,t : some say they dig a hole or graven a safe place

* ' The Salmon,' by A. llussel, 1 864, p. 32.
SECOND SEKIES.— ZOOLOGY, VOL. II.

418 MR. F. DAT ON THE BREEDING OF SALMON FROM PARENTS

in the gravel and there place their eggs or spawn, after the milter has done his natural
office, and then hide it most cunningly and cover it over with gravel and stones.
' Kippers ' have bony gristle growing out of their lower jaws, and may live one year from
the sea, but pine and die the second year. Little Salmons called ' Skeggers,' which
abound in many rivers, are bred by such sick Salmons that might not go to the sea, and
though they abound they never thrive to any considerable bigness."

Willughby *, quoting a communication to Gesner, tells us that generally about the
end of November Salmon ascend for breeding purposes up rivers to their ailluents, where
the eggs are deposited and the young born ; these latter are termed " samlets," while the
old fish descend to the sea.

Ray t observed that Salmon are born in the rivers, whence they descend to the sea.

Pontoppidan (' Natural History of Norway,' 1755, Chapter vi. p. 131) remarks that
Willughby "also confutes Gesner's opinion, concerning the Salmon's breeding in the
sea: he thinks that is done in fresh water, from whence they afterwards go to the sea;
but in this he is certainly mistaken. The Salmon unquestionably breeds in the sea,
though it is not entirely to be denied that they may sometimes breed in rivers also,
for they are found in the midst of Germany, and upper parts of the Rhine, about Basel;
but we are very well assured that the Salmon chiefly ejects its roe at the mouth of rivers,
where they empty themselves into the sea, or a little way beyond, in the salt water, in
this manner: they bend themselves crooked, in order to eject the roe at an aperture
under the belly, and in the mean time they stick their heads down in the sand, that they
may have the more strength. The male comes presently after, to keep off other fish
from devouring the roe, and he then bends his head towards the tail, and ejects his
sperm upon the roe."

Passing over the numerous authors who have merely reproduced the remarks of those
who have preceded thein, we are told by Yarrell J that about the end of 1830 water was
first turned into a certain pond 3 or 4 acres in extent, situated in Scotland, and in
April 1831 one or two dozens of small Salmon fry, 3 or 4 inches long, were taken out of
the river and turned in. In 1833 the first fishing was allowed and several Salmon were
taken with the fly from 2 lbs. to 3 lbs. in weight ; all were perfectly well shaped and filled
up, of the best salmon-colour outside, the flesh well flavoured and well coloured, though
a little paler than that of new run fish. The same author § remarked that " a knowledge

of the growth of young Salmon in a freshwater lake m^y be useful to those

gentlemen who possess lakes near Salmon rivers, from which they can supjfiy them with
Pinks : whether the Salmon thus prevented going to salt water will still retain sufficient
constitutional powers to mature their roe, and by depositing it in the usual manner, as
far as circumstances permit, produce their species, would be a subject worthy of further
investigation."

Knox observes (Proc. Linn. Soc. ii. p. 358, Dec. 19, 1854) that "From the time the Salmon
enters the fresh water it ceases to feed, properly speaking, although it may occasionally
rise to a fly or be tempted to attack a worm or a minnow, in accordance seemingly with

* 'De Historia Piscium,' 1686. t ' Synopsis Methodica Piscium,' 17K'>, p. 63.

I ' British Fishes' (Edition 2), ii. p. 17. § L. c. (Ed. 2), ii. p. 10.

WHICH HAVE NEVEE DESCENDED TO THE SEA. 449

ts original habits as a smolt. But after first descending to the ocean and tasting its
murine food, it never again resorts to its infantile food as a constant mode of nourishment.
He further says, "The absence of this [marine] kind of food forms an insurmountable
obstacle to the preservation of Salmon and of some kinds of Sea-Trout in freshwater lakes."
It has been suggested that a sojourn in fresh water, even should such be only temporary,
may o-ive the first impetus towards the seasonal development of the sexual organs.

Shaw remarked that solitary instances have occurred of large female parrs having been
found in Salmon rivers with the roe considerably developed, and he ascertained that by
detaining the female smolt in fresh water until the end of the third winter, indi-
viduals are found in this comparatively mature condition. Davy* says :-" I have
examined hundreds of parrs .... Male parrs I have frequently found with mature milt,
but never a female with roe correspondingly developed ; on the contrary, in the female
fish without exception, the ovaries have been so small, tbat had they not been sought
after carefully they would have escaped notice." Russel t asks of Mr. Young, « Did he
ever see a female parr with a developed roe ? He never did and never will " $

In this search for roe developed in young Salmon that have not descended to the sea
experiments would appear to have ceased in this country; but after an interval o
about 45 years it was intended to carry them on at Howietoun. Still as it would
not be improbable that some objection would be raised to young fish for this purpose
being removed from rivers to ponds, it was determined to rear the parr from the eggs and
see if smolts or grilse when thus raised in suitable ponds and properly cared for would or
would not breed in fresh waters.

Brown § observed that " no female parr has yet been discovered with roe developed.
This I think, must be taken to read that no young female Salmon in the parr livery that
has not been to the sea had been discovered with ova fully developed. For as I shaU
presently endeavour to show, parr or rather smolt with well-marked parr-bands have given
ally developed ova at Howietoun. Brown || determined to endeavour to rear the ova up
to the smolt stage, and obtained certain ponds, into one of which the sea ebbed and flowed
at Stonehaven in Kincardineshire. They consisted of two ponds, a freshwater one 40 X 20
yards and a saltwater one 60 x 30 yards. The distance between the ponds was about
30 yards with a fall of eight feet. An iron pipe 18 inches in diameter communicated
with the sea. Spawning fish were put into the freshwater pond and afterwards made
their way to the sea-pond, but unfortunately poachers destroyed the experiment, which

was not renewed. , .. . „, .

Clean Salmon having only a thread of milt or roe in them are found m Salmon rivers
during December, January, and February. These would seem to be temporarily sterile f
became ova and milt may be detected by means of the microscope^ Mr P I Well
.ave evidence that these fish do not spawn until the November or December following,
° * . Transactions of the Royal Society of Edinburgh • (Dec. 1854), vol. xxi p. 253 t < The Salmon,' 1804, p. 30.
♦ Tarrell is said to have adduced two instances of parr containing ova m the month of March.
+ , i -looo on 'I Op. tit. p. lib,

8 < Sfcormontfield Experiments,' 1862, p. 89. 1 , , !«.*«,- * -r*

f Siebold has found that certain individuals are not sexually developed; and he supposed that th,s sen. ty
extended throughout their existence. Widegren and others have opposed tins v.ew, centering that such sterility
is temporary, merely affecting a few individuals.

450 ME. P. DAY ON THE BREEDING OF SALMON FROM PARENTS

remaining for 12 months in the fresh water, their ova developing until they are ready to
breed ; and although the fish are discoloured, due to their residence, are very good eating.
Brown caught one of these fish in a Sutherlandsliire river and gave the same report *.

Many have held the opinions which Rasch so fully detailed in 1866 when he
commenced by the inquiry why it is that all the male iish, including those that have
gone to the sea and those that remain behind in fresh water, have their reproductive
organs fully developed, while the female is under the necessity of making the journey to
the sea before being able to spawn. He considered that the female in order to reach
maturity requires far more nourishment than the male, the formation of eggs necessitating
more materials than that of the milt ; that in the ovary of the female the eggs arc
formed nearly simultaneously, and their development is uniform, one being enveloped by
as large an amount of albumen as the other. But in order to produce this albumen a
far greater amount of nourishment is needed than the little female fish can possibly
procure in fresh water — enough for the formation of the embryo, but not sufficient to
provide the nourishment necessary to its proper development. He considered that il'
smolts were prevented going to the sea. they would readily accustom themselves to a
freshwater home, if the piece of water were sufficiently extensive; and should the
water be a very large lake, such as Ladoga, Weneru, or Peipus, and as rich in nourishing
food, the freshwater Salmon will then attain about the same size as the Salmon of the
sea. He also observed that Eetting hatched out numbers of Salmon ova, which he
subsequently turned loose in the Tyri-iiord : and during the two years alter, fish had been
caught in that lake resembling in every respect Salmon proper.

"The question whether any of the migratory species [of the genus Salmo] can be
retained in fresh water, and finally accommodate themselves to a permanent sojourn
therein, must be negatived for the present. Several instances of successful experiments
made for this purpose have been brought forward; but all these accounts are open to
serious doubts t, inasmuch as they do not afford us sufficient proof that- the young iish
introduced into ponds were really young migratory Salmonoids, or that the full-grown
specimens were identical with those introduced, and not hybrids or non-migratory Trout
of a somewhat altered appearance in consequence of the change of their locality. We
have seen the experiment tried at two places in South Wales, by the Rev. Augustus
Morgan and W. Pell, Esq., of Taliaris, and in both cases the Salmon and the pure
Sewin J died when not allowed to return to the sea. However, the latter gentlemau
pointed out to me that the hybrid fishes from the Sewin and the Trout § survived the
experiment, and continue to grow in a pond perfectly shut up from communication with
the sea. In that locality neither these hybrids nor the trout spawn." (Giinther, Catal. of
Fishes, vi. I860, p. 9.) Without alteration, the above passage has been introduced by the
same author into his ' Introduction to the Study of Eishes ' (1880, p. 039).

* ' Stormoutfleld Experiments,' 1S54, pp. 94, 95.

t Lloyd (• Field-Sports North of Europe,' i. p. 301) remarks that near Katrineberg there is a valuable fishen
tor Salmon, 11,000 or 12,000 of these fish being taken annually. They are bred in the lake ; and, in consequence
of cataracts, cannot have access to the sea. They are small in size and inferior in flavour. The year 1S20 furnished
21,817. The question is open to discussion whether these fish are true Salmo solar.

X Salmo trutta. § Salmo far, v.

WHICH HAVE NEVER DESCENDED TO THE SEA. 451

Much of the foregoing seems to me simply hearsay evidence, and several points require
elucidation. First, did any stream flow into these ponds ? and, if so, of what size ? or were
they more or less stagnant pieces of water ? Secondly, did suitable food exist in hese ponds,
or were the fishes fed ? Thirdly, how was it ascertained that no spawn was deposited

Respecting this last question, a very interesting observation was made by Mr. 0. L.
Jackson, Honorary Naturalist to the Southport Aquarium (« Land and Water, June 10th,
187G) that the « Salmon-trout cast their ova in the salt water in the Southport Aquarium
without assuming the appearance of kelts or even leaving off feeding greedily on
shrimps. They did not attempt to make a bed, and the spawn was immediately eaten

bv their fellows." . „ , ,

It will now be necessary to follow out briefly the different stages of growth of the*
fish whieh occurred during the several changes from parr to smelts or grilse. Dr. J.
Davy says' « that parr put into a pond from whieh they could not escape, have grown
in two or three years to he about half a pound in weight, and when caught were
found to be in excellent condition, end still retaining their original transverse markings.
He also (I.e. footnote p. 220) remarks :-« The milt of the young fish, so far as
have observed, is always shed before the parr becomes a smot." It seems cvifleu
from the foregoing that by parr he did not intend to include silvery smelts with pan-
bands ; and if such was bis meaning, it is not in accordance with the changes shown by the
parrs and smelts at Howictoun, where ttsh in the smelt t livery gave milt; and the seme
series of experiments have shown that individual variations in size among these
artificially reared, and under identical conditions, can neither be due to the age of the
parents nor the character of the stock from which the ova or milt has been procured
P A Committee of the Tay Proprietors, on May 2nd, IW, was held at the S —he d
ponds, •■ to consider the expediency of detaining the fry [winch had been hatched Mare
31st, 851, end were 3 or I inches in length] for another year or allowing then to
depart A comparison with the undoubted smelts of the river then descending seaw aids
with the fry in the ponds, led to the conclusion that the latter were net yet smolts, and
ought to ie detained. Seventeen days afterwards, viz. on the 19th May a second
mental was held, in consequence of the great numbers of the fry having in he interim
Is umeu "he mig atory dress. On inspection, it was found that a considerable portion
w Tactual smolts, and the Committee came to the determination to allow them .to depart
Accordingly, the sluice communicating with the Tay was opened, and every am My to,
ITS" afforc led. Contrary to expectation, none of the fry manifested any mclmataon to
eav 1 fpond until the 2«h of May, when the larger ami more mature of the smolts, after
having held themselves detached from the others for several flays went off in a body.
II, -reset similar emigrations took place, until fully one half of the fcy had left the
poifl and deseenflefl the' sluice to the Tay. ... as the shoals successively left the pouch
ibmit one in every hundred was marked by the abscission ot the second dorsal flu A
Later number were marked on the 29th of May than on any other day, in all about

* "Daw ' Physiological Researches,' p. 221.

t S difficult in L, following description of the Howietoun experiments to separate the t-»™^» *£
as these fish do not descend to the sea. on the first return from winch, m a normal condition, the, would be gnlse.
Anyhow, after the deposition of ova they may be termed grilse.

452 ME. F. DAT ON THE BREEDING OF SALMON FROM PARENTS

1200 or 1300 . . . "Within two months of the date of their liberation, viz. between the 29th
of May and 31st of July, twenty-two of the young fish so marked when in the state of
smolts, on their way to the sea, have been, in their returning migration up the river,
recaptured, and carefully examined. This fact may be considered as still further
established, by observing the increased weight, according to date, of the grilse caught
and examined : those taken first weighing 5 to (.H lbs., then increasing progressively to
7 and 8 lbs., whilst the one captured 31st July weighed no less than 9| pounds. In all
these fish the wound caused by marking was covered with skin, and in some a coating of
scales had formed over I lie part " *. This question of smolts returning or not the same,
year as they migrated to the sea, I do not propose entering on here.

llespecting one half of the purrs migrating seaward one season, and the remainder not
doing so until the next year, ltu>sel observes, ".V new hypothesis was brought out, to
the effect that the females descend the first year and the males the second " (op. cit. p. 17).
But on investigation it appeared "thai the fish remaining during the second year
consisted of both males and females, the milt of the males being fully developed, while
the roe of the females was discernible only by a microscope " (/. c. p. 18).

Mr. Dunbar, who annually hatches about 500,000 Salmon ova in the Thurso river in the
county of Caithness, informed -Mr. Young that about 8 per cent, became smolts at the
end of the first yearf, and about GO per cent, al the end of the second year, and the
remainder, or 32 per cent., at the end of the third year. The fish which were hatched
in the spring of 1837 at Drumlanrig did not assume the migratory livery and seek to
depart in May 1838 as Mr. Shaw expected, but did so in May 1839 %. It has been
argued that young Salmon may have their growth stunted and their instinct overruled by
being kept in a state of comparative confinement.

" Until the parr takes on the smolt scales it shows no inclination to leave the fresh
water. They cannot live in salt water. This fact was put to the test at the ponds, by
placing some parrs into salt water — the water being brought fresh from the sea at
Carnoustie — and immediately on being immersed in it, the fish appeared distressed, the
tins standing stiff out, the parr marks becoming a brilliant ultramarine colour and the
belly and sides of a bright orange. The water was often renewed, hut they all died — the
last that died lived nearly five hours. After being an hour in the salt water, they
appeared very weak and unable to rise from the bottom of the vessel which contained
them, the body of the fish swelling to a considerable extent. This change of colour in

* Report of Committee on Storraontfield Ponds to British Association, 1S50, p. 4-V>.

t I conclude he meant to say at the termination of the first year after being hatched, for they do not descend
seawards in their first year.

t " It was said of Mr. Shaw's experiments that the two-years' freshwater residence of the fry was ascribable to the
•difference of temperature between the waters of the Nith from which the ova were taken, and the waters of the
ponds in which they were hatched and reared.' Put where is the evidence as to what was the difference in temper-
ature, or whether there was any at all? On inspecting Mr, Shaw's Observations, for information on this point, we
can only find that the temperature of the ponds, as compared with that of the river, was on one occasion three
degrees below and on another six degrees above. So whatever difference there was, seems to have been in favour
of the ponds stimulating, not retarding, as compared with the river or natural habitat. Put suppose it were other-
wise, what then '? Wo know that a lower temperature might retard the hatching of the fry by a week or two, or
their growth by half an inch or half an ounce ; but we have no ground for supposing that it would retard for a
whole year such a change as that of assuming the migratory dress." (Pussel, I.e. p. -ts.)

WHICH HAVE NEVER DESCENDED TO THE SEA. 453

the fish could not be attributed to the colour of the vessel which held them, for on
being taken out they still retained the same brilliant colours " *.

In the Brighton Aquarium about 20 small Salmon parr were received from Mr.
Berrin-ton, the Chairman of the Usk Board, and these were placed in fresh water, where
they soon began to feed and took to their tank. In the succeeding May, about eight months
after they had been received, most of them commenced to assume the smolt livery, but
four remained as parrs. Salt water was gradually introduced ; but this did not prove
fatal to the parrs, as it was feared it might, while the smolts became quite rampant with
pleasure as the water grew more and more salt, until at last no fresh water was left
and it became purely salt, Then the parrs, which had remained parrs up to this time,
be-an to assume the smolt livery, and the change is described as truly marvellous. They
ate five times as much as previously, were in incessant and rapid motion all day, and
their growth became perfectly astonishing f.

Bertram remarked of the Stormontfield smolts, in 1861, "One fish which had been
detained for three years for the purpose of discovering whether the species will grow in
fresh water without being permitted to visit the sea, was found to be fully twice the size
of the largest smolt " J.

The Duke of Buccleuch's gamekeeper at Bowhill was for some years in the practice
of putting a few smolts into a freshwater pond, and feeding them regularly with bullock's
liver He reported that the smolts which grew into Salmon throve for about three
years and then died, but that Bull-trout smolts kept in good condition for a longer

period. .

« Into a deserted stone-quarry near Coldstream, filled with rain-water, two smolts, about
3 inches long, were put by boys out of mere amusement. One of these grew into a
Salmon, which when 5 years old, weighing H lb., was caught and sent to Mr. Stoddart
of Kelso who had it boiled for dinner. He reported that it was not unpalatable. Ihe
other smolt grew into a Bull-trout and lived for seven years. It died during a very
severe winter when, on account of the water being frozen, it could not be fed " §.

« Shaw has demonstrated, in the most conclusive manner, that those small Salmonoids

which are generally called Parr, are the offspring of the Salmon, and that many males

from 7 to 8 inches long have their sexual organs fully developed, and that their milt

lias all the impregnating properties of the seminal fluid of a much older and larger fish.

No Parr has ever been found with mature ova" ||.

'"In the Exe, parr-marked fish (graveling or smolt so-called) have been taken with
ova actually exuding from the fish, on and previously to its being handled to remove the
hook Some were sent to me by Mr. England, fishing in the Exe, some four years since,
and preserved in glycerine, most certainly as a novelty, since which many such cases
have come before me " (Frank Gosden, < Land and Water ').

I think I can adduce conclusive evidence so as to remove any doubt respecting smolts
m-ilse giving ova without descending to the sea, as they have done so in the

or gJ

* Brown, ' Stormontfield Experiments,' pp. 61, 62.

t Francis Francis, ' Field,' 1S79. t ' Harvest of the Sea,' p. 110.

§ Extract from ' Eeport of Experimental Committee to Tweed Commissioners.'

|| Giinther, ' Introduction to the Study of Fish,' 18S0, p. 639.

454 MR. F. DAT ON THE BREEDING OF SALMON FROM PARENTS

Howietoun establishment. But the assertion that the milt of parrs has all the impreg-
nating properties of the seminal fluid of a mueh older and larger fish must he received
with a qualification, at least if the foi'ogoinn' is intended to signify that it always has
such.

It is frequently stated to he the rule that Salmon deteriorate on entering rivers ; but
this view must be modified by observing that such is also, partially at least, due to their
being mostly in a breeding condition, at which period they would naturally fall off in
condition, oven were they to remain in the sea. At other times some ascend without
deterioration, demonstrating such to lie possible, owing to the state of health or
condition which they were then in.

Shaw mentions * that in January 1837, he took a female Salmon, weighing 14 lbs., from
the spawning-bed, Prom whence he also look a male parr weighing one and a half ounce,
with the milt of which he impregnated a quantity of her ova and placed it in a stream
connected with a pond, where, to his great astonishment, the process succeeded in every
respect as it had done with that which had been impregnated by the adult male Salmon,
and exhibited, from the first visible appearance of the embryo fish, up to their assuming
their migratory dress, the utmost health and vigour. In January 1S3S, he took
another female Salmon weighing 1 1- lbs. and two male parrs from the same spawning-
bed and impregnated two lots of parr-ova with the milt from the two parrs, and afterwards
placed them in two different streams enclosed in boxes, open at the top, temperature
1") . In December 1838 he took a female Salmon from the river weighing 11 lbs., and
lour male parrs from the same spawning-bed. After impregnating four different lots of
ova, one lot to each individual pan-, he placed the four parrs in a pond, where they
remained until the following May, at which period they assumed the migratory dress.
The ova were placed in streams to which no other fish had access, and where they became
mature in a similarly progressive manner to those already detailed, thus clearly demon-
strating that the young Salmon of 18 months old t, while yet in the parr or early state,
actually performs the duties of a male parent before (putting the river. While the
males of the three several broods which occupied ponds No. 1, 2, 3 continued in a breeding
state, which lasted throughout the whole of the winter of 1838-39, he impregnated the
ova of three adult female Salmon from the river with the milt of a male taken from each
of the three ponds, and the whole of these ova matured. This, he deemed at once,
removed any doubt which may have been entertained regarding the constitutional
strength of individuals reared under such circumstances.

At Stormontfield, in the season of 1857 (November or December), milt from a parr was
used to fertilize the ova of a 164bs. Salmon, and in 1858 they had fry of Salmon, fry of
grilse and Salmon, fry of grilse, and the fry of the Salmon and parr. " On the closest
inspection, no difference was perceptible cither in the form, colour, size, or markings
of any of these fish. There were larger and smaller fry to be seen amongst all these
hatchings " (Brown, ' Stormontfield Experiments,' p. 74).

* 'Transactions of Royal Society of Edinburgh,' siv. p. 561,

t As young Salmon hatch at the latest in March, these young fish must have been at least 21 months of age, if in
their second season.

WHICH HAVE NEVER DESCENDED TO THE SEA. 455

At the same establishment a suggestion which had been advanced— that the anomaly
rf some parrs migrating at the end of one year after hatching, and the remainder
tfter two seasons, must have been caused by the first being the produce of Salmon,
vnd the second of grilse-was clearly refuted. The fry, which numbered about 200 000,
*rere solely the produce of 19 male and 31 female Salmon, spawned in 1859 ; and ot these
some remained in the ponds as parr, while others migrated seawards as silvery smo Its.
These same experiments showed that marked grilse of one year return as Salmon the
next • that all the smolts of one year do not return the same year as grilse, the one hall
returning next spring and summer as small Salmon (I. c. pp. 92, 93) ; also that marked
smolts w°ere caught as grilse the first season and Salmon the second (I. c. p. 94).

Asain, some authors have held that were Salmon able to migrate into large freshwater
lakes, where a sufficiency of suitable food existed, they would be able to return to the
streams where they had been reared as well-developed Salmon, a theory denied by
others, who assert that Salmon never increase in weight in fresh water ; but it this were
so, how do they grow at Howietoun ? #

Dr Murie (Proc. Zool. Soc. 1868, p. 219) gave an account of some Rhme Salmon
(Salmo salar), hatched in the Gardens of the Zoological Society, in February 1863 and
which lived in fresh water up to 1868. In May 1805, at the usual period of migration,
those which were smolts tried to leap out of the tank; similar phenomena occurred m
1866 and "it was further observed that those which had assumed the silvery dress in
the beginning of the year, again lost it in the autumn and became distinctly parr-
marked " In this autumn a good many died, some in the smolt, others m he parr-
livery ; and 1867 was a repetition of 1866. During the first and second years t he _ young
fish seemed to grow, and attained a length of from 3 to 6 inches During he third and
fourth year they grew much slower, and attained to from 5 to 7 inches m length ; onl>
I lived to the'fifth year. These fish were kept in a small tank with a running
stream of fresh water, a fact which is very necessary to bear m mmd, because, although
the conclusion drawn that the fishes retained in fresh water were subject to an arrest ot
growth, such may have been due to the small space in which they were confined, or
the limited amount of water which they obtained, irrespective of which, the changes ot
temperature may have been great.

Yarrell ('Growth of Salmon in Fresh Water ') remarks that the view that the rate ot
growth in young Salmon has some reference to the size of the place to which they are
restricted, receives further confirmation in these river, lake, and well specimens. The
molt taken from the well in July 1838, where it had been confined ^^™
rather smaller in size at that time than the smolts in the preceding April, though both
were Pinks of the same year, namely 1837 . The smolt taken from the lake in August ^1838
which then measured 7* inches, had also grown more rapidly than that in the weU but £ad
not acquired the size it would have gained had it been allowed to go to sea Further ,t
Lay be observed, that the Salmon peal from the lake in August 1837, then 8 mon hs o d,
Thou h perfect in colour, is small for its age ; while that of July 1838 or 29 months old
L comparatively still more deficient in growth, supposing both fish to have resulted
Lm Xks of the year 1836, and been put into the lake at the same time ; ot which there

SECOND SERIES.— ZOOLOGY, VOL. II.

456 ME. F. DAY ON THE BEEEDING OF SALMON FEOM PAEENTS

was no doubt, since the lake, the formation of which, though commenced in the autumn
of 1835, was not finished till February 1836, soon after which the first Pinks were put in."
Before detailing the experiments I would wish to draw especial attention to the
necessity of complete segregation of the fishes experimented upon, and the produce
reared. Crowding many forms into one tank, and from this crowd selecting various
varieties or so-called species simply from their appearance, and hybridizing from them,
is likely to occasion irreparable errors and do more injury to investigation than if the
subject were left quite alone.

I do not propose giving a detailed account of the Eowietoun ponds, how care is taken
that no one can have unauthorized access there, nor how the segregation carried on is
most complete, as I have reason to believe that a distinct paper on this interesting
subject may l>e expected, the contents of which I should not wish to anticipate.

In December 1880, Sir J. II. Gibson-Maitland, F.L.S., obtained some Salmon in the Teith,
from which he procured eggs and milt, and they hatched at llowietoun in March 1881.
In due course these young fish were transferred to pond no. 7 at llowietoun, which has
its sides and bottom lined with wood, is 100x15 feet in surface extent, 8 feet deep in
the centre and (>l at either side, while the average supply passing through it is a
million and a half gallons of water a day.

In July ]883, in the presence of the foreign and other Commissioners to the Great
International Fisheries Exhibition, this pond was netted, when it was seen that some of
the Salmon-fry, then 2 years and I months old. were of a general golden colour, spotted,
and in the banded parr-stage livery, while others were beautiful silvery smolts, similar
to such as w^e generally find descending to the salt water, and in which, in certain lights,
parr-bauds were visible. Although, as a rule, the smolts were the largest, still some few
of the parrs exceeded smolts in size. All present were convinced that these fish were
the young of the Salmon, and in excellent condition.

October 10, 1883. One of these fish under a pound in weight, in the smolt livery and
showing parr-bands in certain lights, jumped out of the pond and was found dead on
the path. On being opened it proved to be a female with comparatively large ova,
which were of a deep reddish colour, thus almost disposing of the statement already
alluded to that "no parr has ever been found with mature ova." I say "almost,"
because as yet the ova were not quite mature, although they would have become so had
not the fish met with an untimely fate ; also the objection might be raised that it was a
smolt, not a parr. Now although the parr-bands were present certainly the golden tinge
-ecu in the young of the Salmon was absent. Up to the end of November 1883 several
of these fish, during the night or early morning, had jumped out of the pond, and it was
surmised that some might have been carried off by vermin, as birds or rats, in the early
hours. On the 29th two more, one being Hi inches in length, were found dead, and
some vermin had devoured the lower part of its head. It was of a silvery smolt colour
with parr-hands. On being opened it proved to be a male full of ripe milt *. A net
having been put into the pond three fish were examined. The first was a silvery smolt,

* For descriptions of the specimens see Proc. Zool. Soc. 1884. p. 2\.

WHICH HAVE NEVER DESCENDED TO THE SEA. 457

.0-8 inches in length, which in certain lights distinctly showed ten parr-bands, one
dack spot on the preopercle, three on the opercle, while the upper two thirds or half of
he body was spotted with black, as was likewise the lower half of the dorsal fin,
vhich had a white anterior and outer edge. There were some red spots on the body.
[n this fish the generative organs were not developed. The second was more distinctly
parr-marked, but otherwise very similar ; it was 9 inches in length. The third was in the
parr-livery with 10 well-marked parr-bands on either side and three black spots on the
opercle ; it was 0-3 inches in length and a male full of milt. As a rule the females in
this pond were silvery, with ova which were not quite ripe, while the males on the
contrarv, being pressed, gave ripe niilt. As to colours the fish seemed to be more spotted
than is 'seen in specimens from rivers *, while their bodies were well formed although
not of a large size, the largest being 13* inches long. Those in the golden parr livery
appeared to be males, while the silvery smolts were generally females ; in all the finger-
marks were visible. These fish rendered it very evident that the relative growth of the fry
does not depend on the size of pond, quality, quantity, or variety of food, or amount
of water, as all had been treated alike, yet they varied in length from four inches to

thirteen and a half.

During the month of May 1884 sixteen of these fishes jumped out of the pond and
were found dead, and the breeding-organs in all were found to be developing in a
satisfactory manner.

August 28, 1881. An examination was made of pond no. 7 and a smolt li lb. in
weight was removed f. On being opened it proved to be a female with its ova well
forward, the largest being Ol of an inch in diameter; the total length of the fish was

14j*1 iuclics.

October 3rd. A smolt 13 inches in length, which had jumped out of pond no. 7, or had
been dragged out by vermin, also proved to be a female, and the eggs were 0-2 of an

inch in diameter.

Now some of these fishes tried repeatedly to jump out of the pond $ and these generally
met with fatal injuries ; they sprung out towards the upper end near where the water
entered as if wishing to ascend the stream. Wire netting one foot in height was now
fixed around the upper half of the pond, which prevented the fish springing over the
bank About the middle of October, fungus (Saprolegnia ferax) broke out among
these fish, possibly due to injuries occasioned by striking themselves against the wire
nettino- It was also observed that they showed signs of ceasing from feeding, so a
number of small Lochleven Trout were added to vary the diet, but they did not appear to
consume many of them.

* The beautiful illustrations of grilse in Sir W. Jardine's magnificent < Salmonid* ' do not resemble in colour the
Howietoun fish, which latter are more spotted than his 2 lbs. 13 oz. specimen from the Solway F.rth, taken m July,
and even more so than his 3§ lbs. captured in August.

f Described in Proc. Zool. Soc. 1SS4, p. 5S1. •

% Brown CStormontfield Experiments') observes that he « in the month of February 1838 caught a dozen and ahalf
parrs in the Tav, and kept them confined in a stream of running water, and by the month of May the whole of them
had become smolts; but some had leaped out of their confinement in their struggle to find the. way to the sea. and

■were found dead on the side of the pond" (pp. 7 and S).

Ob*

458 MR. F. DAT ON THE BREEDING OF SALMON FROM PARENTS

November 7, 1884. A suiolt, or more properly a grilse, 1| lb. weight, was found lying
almost dead by the side of pond no. 7, and from it about one hundred apparently ripe eggs
were taken, and milted from a Lochleven Trout. On January 23, 1885, eighteen hatched,
and when I saw them on February 10 they looked remarkably well and vigorous, while
there did not appear to be any deformities among them. So far as I am aware, this is the
first successful attempt in ( rreat Britain at breeding from Salmon which have undoubtedly
passed their entire existence in fresh water, while it is likewise interesting as a case of
hybridization, and in a form which, so far, at llowietoun had been productive of sterile
offspring.

November 11, 188 !•. About 12,000 eggs of the Lochleven Trout were fecundated from
the milt of a grilse, and were placed in hatching-tray no. 1. They did very well, only 87
dead ova being picked out, and hatched on January 28, 1885.

November 14, 1884. The water was let oil' from pond no. 7, in which the young
Salmon had been kept, as it required being re-charred this year, while the contained fish
were taken to pond no. 5, just re-done up, and previously inhabited by Brook-Trout, but
which were now turned into the burn, as they were too old to be worth further
preserving. Sixty-eight young Salmon were transferred, the majority being males,
while these fishes showed all the gradations in colouring from the golden and banded
parr to the silvery smolts, these latter, however, not having lost their parr-bands, while
some of the large smolts or grilse were distended with ova. All the fish in the full parr-
livery were males, but smolts and grilse in some instances also had milt. About two dozen
eggs were obtained IVom one of the females, but they did not germinate, probably not
being quite ripe. Three smolts which were too much injured to live, due to their
persistent attempts to escape, were opened, and the ovisacs were found distended with
eggs, almost but not quite ripe, being still slightly adherent to the ovisac.

As these youug Salmon hecame ready for continuing their species, in December the
following were obtained and laid down in the hatching-bouse : —

December 1, 1881. 1500 eggs from two smolts or grilse which were treated with the
milt of one of the males. These eggs averaged each 0"22 of an inch in diameter, and
Avere placed in hatching-tray no. 101 a : about 400 of these eggs batched on Feb. 21, 1885.

December 9, 1884. About 4000 ova of these grilse were impregnated with the milt of
one of the males. These eggs were each 0'20 of an inch in diameter and were placed in
hatching-tray no. 108 a : about 2200 of these eggs hatched on Feb. 27, 1885.

On the same day 400 ova of these grilse were treated with the milt of a Lochleven
Trout. These eggs were each 0*22 of an inch in diameter, and were placed in hatching-
tray 108 b. The milt appeared to be strong enough where it acted, but only appeared to
do so in half the eggs.

December 13, 1884. About 2500 eggs of the grilse were milted from the same lot in
the pond, and placed in hatching-tray 84 c. The diameter of each egg was 0-22 of an
inch, and about 200 of these eggs hatched on March 5.

At nearly the end of December the remainder or 48 young Salmon which had been
bred from, were removed from the wooden-lined pond no. 5 into the earthen pond
no. 10, it being of larger size, and none jumped out.

WHICH HAVE NEVER DESCENDED TO THE SEA. 459

I have already shown that the fact of parr having milt prior to descending seawards

iaS been generally recognized, but that smolts or grilse kept under similar conditions

mve been almost always on the contrary declared not to give ova. But now we ha.

he experiments made at Howietoun to consider, and these have undoubtedly proved

hat young male and female Salmon kept in suitable freshwater ponds may develop

M or ova. These fish were hatched in March 1881, and most of the male parr had milt

in November 1883, or when 2 years and 8 months old; also two or three smolts of the

.amea^e had ova which probably would have matured, but they jumped out of the pond

nTso met with their deaths. In November or December 1881, or at 3 years and 8

f 9 Tenths' age, all these fishes seemed to be ready to breed, and it appears

obabTethen that' Salmon-parr do not possess milt for the purpose o nupregn. mg

the ova of the old Salmon, or "to mingle with the River-Trout, a Dr Gunthe

su^ests but to be ready to fertilize the eggs of any ascending grilse, for in some

Sties' the"! latter deposit their ova prior to the usual period when adult male Salmon

are present for breeding purposes.

S„me authors bold that Salmon axe freshwater forms that proceed to , the sea, and that
there is a physiological necessity for their doing so in order to be able to .produce *
Other,, who I believe are more correct, consider these fish marine to - U a ascend
into fresh water to deposit their eggs and rear their young. As rt , eviden f« ■*
Howietoun experiments that these fish con give eggs without descending to salt water
fu i tToses'of the " physiological necessity- theory Possibly t bis imgra ion ma
be classed among the instances of "inherited instinct" and may *•«»»*. -
of Salmon as have been reared for a few generations in fresh water. I -may be that
in some such manner the "laud-locked Salmon" ot various "T^TL^
originally differentiated from their fellows, an interesting instance of which I propose

t0 „llh O^e in 1S81 too, about 100 Sea-trout and Salmon smolts, which he

turned into Loin* Ash, Co. Tyrone, which bas no access to the sea. April SO, 1SS3 he

luir a gX 14 inches long in this lake, where Salmon had not previously been

en Its ablomen was so distended that be considered it would have spawned very

I^more especially as it was taken, £. outb of tbeon, stream that enteis

Z netl ex^Tl TjT^t^L ■"*» «~ ^ ^
S« iu "measured by me were each 0-25 of an inch in dieter. Proved
t, re'bad been a ripe male iu the vicinity, there seems no reason why this fish u
not have deposited her eggs, and such would thus have given rise to a land

''""On the approach of autumn," observes Shaw (Trans Roy. Soc Edin ^ P. M*
footnote) "the whole of the Salmonids, resident as well as migratory, while in fresh
wato acquire a dusky exterior, accompanied by a considerable increase of mucus or

* Described Tree. Zool. Soc. 1S84, p. 584.

460 ME. F. DAY ON THE BREEDING OF SALMON FEOM PARENTS

slime. The fins also become more muscular. However, on the return of spring, they
resume their wonted beautiful silvery colouring, aud the fins, the cartilaginous portions
of which are frequently damaged during the winter floods, grow up and acquire their
former outline."

Different authors have assigned various reasons why a parr on becoming a smolt (which
it does, as a rule, when commencing its seaward journey) assumes a silvery livery. Davy *
suggested " that the young remain in fresh water until they have acquired not only a
certain size and strength, but also additional scales, fitting them, in their smolt stage, to
endure without injury the contact of the saline medium." Couch demurred to such an
opinion in 1866, observing that the silvery colour of smolts is not due to their acquiring
additional scales, but owing to a deposit of bright soft matter which shines through the
transparent scales. Dr. Giinther f remarks, respecting the River-Trout, that they " fre-
quently retain the parr-marks all their lifetime ; at certain seasons a new coat of scales
overlays the parr-marks, rendering them invisible for a time, but they reappear in time,
or are distinct as soon as the scales are removed. When the Salmones have passed this
parr-stage .... a new coat of scales overlays the parr-marks." But as these fish do not
shed their scales, Dr. Giinther appears to hold to the untenable proposition that an
additional coat of scales is developed on these fishes at certain periods +. It was like-
wise stated in one of the conference papers read at the " Great International Fisheries
Exhibition " in 1883 §, that the young of the true Salmon " do not venture into the sea
till another skin of glistening scales has been formed over their first skin. They then
receive the name of smolts. If put into salt water before getting this silvery dress
they die." I do not propose here entering upon the discussion of these opinions, so
wide spread, but nevertheless entirely erroneous. Any one who will examine a series
of specimens can readily verify for himself that there never is an extra coat of scales
as above stated.

Before concluding this portion of my paper, I may observe that the teeth in the vomer
of a parr generally are as follows : — three on the head of that bone, below or behind
that a pair, next two placed one at an angle to the other, and subsequently eight in a
single row, but with their points somewhat divergent. There is very little change seen
in those of the smolt, while in a grilse 13 inches long, that spawned this year, the same
dentition prevails. This raises the question whether the loss of teeth, as generally seen,
is not partly due to the hardness of the food they consume ; this, however, is not the
case in the Howietoun pond, as what they are fed upon is soft.

I omit any detailed description of the grilse, and decline entering upon a lengthened
discussion of the decision of a Committee of the Commissioners of the river Tweed,

* ' Physiological Researches,' 18G3, p. 250.
t ' Introductiou to the Study of Fishes,' 18S0, p. G32.

% See Proceedings of Zoological Society, 1884, p. 22, showing the impossibility of an extra coat of scales being
developed.

§ " Salmon and Salmon Fisheries," by D. Milne Holme. (Conference Papere, p. 4.)

WHICH HAVE NEVER DESCENDED TO TUE SKA. 461

who stated, in their report for I860, as follows :— " Our opinion, from the experience of
the last twenty years, is, that grilse never become Salmon of any stage whatever." As
grilse were reared from Salmon-ova at Ilowietoun, it may well be asked in reply, if
such were a fact, how it is that grilse may he raised from Salmon-eggs. Russel, when
refuting this argument of the Tweed Commissioners, and also of Mr. Mackenzie, remarked
upon the difficulty of obtaining conclusive evidence after the fish has assumed its migratory
habits and can no longer be kept under inspection to see if the grilse becomes a Salmon.
But, as I have now shown, the reverse experiment has been successfully carried out
at Ilowietoun, and eggs of Salmon have been proved to give birth to fishes which in
time become smolts and then grilse.

A difference of opinion respecting the persistent or variable size of the eggs of Salmo-
nidoe appears to prevail, a difference which could not exist if the authors would take tin-
trouble to measure them when freshly removed from the parent fish, or as lying in the
hatching-troughs. For these eggs do not increase in size during the process of incuba-
tion, although such as die become sodden, white, aud somewhat distended. Uniinpreg-
nated eggs do not, as a rule, take on this opaque white aspect, but remain clear
throughout the nesting-season, neither do they augment in size.

I propose offering a few remarks upon whether the eggs of the Salmonidoe, of the
same species, are invariably of the same size, and especially if the age of the parent has
any connection with such differences as maybe perceived; also, if any variations in
the size of the eggs from a single fish are to be seen.

Ilarmer in 1707 * gave a table showing the number of eggs which he had observed in
certain forms of fishes, among which the Salmonidae were not included. He remarked,
" From this table it appears that the size of the eggs is nearly the same in great and
small fishes of the same species, at the same time of the year." And these observations
may be found, with but little variation, in the writings of authors from that period up to
recent years, as perhaps in no branch of biology are assertions once made more adhered
to without re-investigation than in Ichthyology. Passing on to the 'Zoological Record '
of 1864, p. 179, we find that Professor Malmgren, having observed that certain Salmonoids
in a lake in Finland were descendants of the common Salmon, whose access to the sea had
been cut off owing to an elevation of the land, mentioned as one of the present differences
that this freshwater and dwarfed breed gave smaller ova than Salmo salar. In this
conclusion Dr. Giinthcr did not coincide, remarking that " the last character [or size of
the ova] will be considered very significant by all who may have a more extended know-
ledge of fishes, as the size of the ova is not only invariably the same in individuals of
whatever size, but, as far as our experience reaches, is even often characteristic of the
species of a genus." This opinion, slightly modified, Dr. Giinthcr still apparent ly held in
1880, for he asserted t that " the ova of Tcleostean fishes are extremely variable in size,
quite independently of the size of the parent species. The ova of large and small
individuals of the same species, of course, do not differ in size."'

* Transactions of the Royal Society, " On the Fecundity of Fishes."

+ • Introduction to the Study of Fishes,' p. 159.

462 ME. F. DAY ON THE BREEDING OF SALMON FEOM PARENTS

On the other hand, Mr. E. Blanchard in 1S6G* observed that the ova " of the grilse
are always sensibly smaller than those of the adult Salmon." Livingston-Stone, 1877 f,
remarked that in American Trout (Salmo fontinalis) that reside in spring- water, winch is
equivalent to a diminished supply of food, smaller eggs are developed than in such as
reside in brooks. On the M'Cloud river % " it was noted, in 1878, that the parent
Salmon were unusually small, their average weight being under 8 lbs." This small size
was stated to be undoubtedly caused, in whole or in part, by the fishery at the canneries
of the Sacramento, where the 8-inch meshes of the innumerable drift-nets stopped all
the large Salmon, but let all the small ones through. The eggs when taken proved to
be at least one third smaller than those of most previous years, and the average number
of eggs to the fish was about 3500 against 4200 in the previous year. Livingston-
Stone in 1882 §, writing respecting the eggs of Salmo fontinalis, observes that those
from the small fish of the elevated rivulets are not so large as those of the finer breeds
from warmer streams, concluding that the ova of these last are fully twice as large as those
found in the race residing in mountain rivulets. From this period, observations have
been frequently recorded to show that the eggs of the Salmonoids undoubtedly differ in
size consequent upon certain physical causes ; hence the assertion that in this family " the
small size of the ova [in Brachymystax coregonoides, Giinth.] is of sufficient importance
to separate these fishes generically from Salmo" j| may be questionable, especially as the
typical specimens examined in London were only 18 and 21 inches respectively in
length.

The following figures are all taken from examples in the Howietoun fisheries, unless
otherwise noted ; and in order to make the statement plainer, the eggs of the various
species or crosses of these fish are given : —

Sal, ,io salar, 16 lbs. weight. General size of eggs 0-24 of an inch.

„ ? „ „ from 0-25 to 0-30 „

„ small „ „ 0-20

,, „ grilse reared in ponds „ from 0-20 to 0-22 „

„ trutta, small ,, „ ,, 0-175 to 0-18 „

„ levenensis (var.) 8 years old, from one pond, the general size .... 0-24 „

but some are 022, and a few 0-21. and in three or four 0-20 ,,

,, ,, 7 years old. General size 0-19 „

„ 6 „ 0-18 to 0-19

2 and 3

,, fontinalis 4

2

Hybrid S. fontinalis, S. alpinus 2

. . . 0-17
. . 0-18
. . 0-17

. . . 0-14
0-13 to 0-15

* ■ 1'uissons des Eaux Douces de la France,' p. 401. t ' Domesticated Trout,' 3rd edition.

% Report of United States Fish Commission.

^ Bulletin of the United States Fish Commission, ii. 1882, p. 11.

|| Catal. Fishes Brit. Museum, vi. 1860, p. 103.

WHICH HAVE NEVER DESCENDED TO THE SEA. 403

The foregoing are only some out of the many measurements which I have made, but they
tend to show very conclusively that the size of the eggs iu Salmonoids increases with the
age and probably the condition of the parents ; thus the ova of grilse are inferior in size
to those of the Salmon. And in large Trout as in the older Lochlevens there is an
augmentation in the thickness of the shell *, a rather important question to the fish-
culturist, because such shells require a strong embryo to readily hatch, which is
obtained by giving an increased supply of water during their " eyed " period. Whether
it was accidental or otherwise, still the Trout eggs obtained latest in the season were of
rather larger size than such as had been taken in the earlier months.

Next we come to the consideration of whether any difference is perceptible in the eggs
of two parents of the same age that have been kept under identical conditions. Amoug
the young Salmon reared at Howietoun from eggs obtained from the same parents, and
reared in the same pond, some taken on December 1st, and again on the 13th, 18S4,
averaged 0-22 of an inch ; some from similar fish on December 9th averaged O20 of an inch,
while others taken the same day were 022 of an inch. The same phenomenon has been
observed among the Lochleven Trout, wherein, although the diameter of the eggs of fish
at 8 years of age was generally 024 of an inch, in some it was 022, in a few 0'21, and
in three or four O20 ; whilst in six-year-olds it has been observed to vary from 0"1S to 0*19
of an inch. That such variations might theoretically be anticipated most of us must
admit, for whatever causes deficiency of nutrition might occasion diminution in the size
of the ova. Also the size of the parent ought possibly to be taken into account, for
disease, crowding them when young, and other causes may dwarf these fishes, and such
would probably give comparatively small eggs.

Can we detect any variation in the size of the eggs given by a single fish ? I have
already adduced an instance of this among the Lochleven Trout. In November 1884, I
placed 80 unimpregnated eggs of one of these fish in a tube of water, wherein they were
kept for a week, the water being changed daily. At the end of this period I measured
each ovum, and found 75 were 0-25 of an inch in diameter, three were 020 of an inch,
one was 0-175 of an inch, and one 0'15 of an inch. Looking through the hatching-
troughs at Howietoun it appeared that from 4 to 5 per cent, of the eggs are less than the
normal size of the remaining 95 or 96 per cent.

I must here refer to the grounds on which I believe that offspring raised from the large
eggs are superior to such as are reared from the smaller ones. Two batches of Lochleven
Trout were spawned on November 2nd, 1882, the parents of one lot having been hatched
in 1875, and of the other in 1876 ; these ova were similarly treated and the young came
out during January and February 1883. At Craigend are two ponds which have been
constructed for the reception of young Salruonidae, each being 100 feet long and about
25 wide : one is nearly on a level with the other, and the same stream passes through
both. Into these two ponds the two lots of fry were turned, those from the older (1875) or

* The micropyle in the Salmon's eggs has nine pits or depressions around the opening, whereas there are none in
the ova of Trout, including Sea-Trout.

SECOND SERIES. — ZOOLOGY, VOL. II. 69

464 MR F. DAT ON THE BREEDING OF SALMON FROM PARENTS

seven-year-old parents having the lower pond, those from the younger or six-year-old
lish having the upper pond, while both lots were similarly treated. At the end
of November 1883, those in the lower pond were about one fourth larger than those in
the upper pond, and as they received the stream after it had passed through the upper
pond, this gave reason to suppose that the large eggs produced superior fish to those from
the smaller eggs. In 1884 a second experiment was tried in the same two ponds, which
were stocked with young Lochleven Trout, bred from parents of the same age, spawned
the same day, and hatched in the same room. By the end of the year no difference in
size or appearance was perceptible between the two lots. It would thus seem that the
larger eggs given from older parents produce fry which grow faster than do those from
the smaller eggs furnished by younger parent-.

Mr. Francis Francis (Fish Culture, 18<!5, Appendix, p. 309), when alluding to the
Salmon fishery at Doohulla in Ireland, has stated, " A curious fact occurred also, which is
worth noting, showing that big smolts make big lish. Among the smolts marked were
thirteen which had far outstripped the rest in size; and among the grilse were some
which had also considerably outstripped the rest. Most of these smolts were marked.
Among the rest was one very line fellow, which Mr. Ilamsbottom had to hold in both
hands while he was marked. They were cutting off the adipose or soft fin, and the
operator, who held the scissors, remarked as they came to this fish, 'What a pity to
mutilate such a fine fellow ! We will only cut oil' a piece of his I'm.' And the fin was cut
in halves, leaving a small triangular piece. This very lish was caught, and as he was
the largest smolt, so it was the largest grilse and weighed 7.1, lbs."

The immediate and possihly remote result of having eggs from young fish for stocking
purposes is one which requires the earnest consideration of the fish-culturist. Tor I
have already shown that by selection a more rapidly growing race may be reared, and I
now propose examining whether small eggs from young parents may not give bad results
as to the number hatched or the quality of the young.

For several seasons it had been remarked that eggs from young mothers are subject to
a greater percentage of deaths than those taken from older fish. On November 13th,
1884, about 500 eggs having a diameter of 017 of an inch were taken from a rising two-
year-old Lochleven Trout and they were impregnated from another of the same race of
the usual size. Out of these eggs only about a dozen hatched on January 28th, and seven
lived to be turned into the rearing-pond. In this instance the immaturity was on the
mother's side.

On November 29th, 1883, 4500 eggs of the Lochleven Trout (of the season of 1875)
were milted from the parr of a Salmon raised at Howietoun and which had been hatched
in March 1881, and consequently were a little over 2 J years of age. The mortality of
these ova was only about 2 per cent. ; but such by no means gave a true index of the experi-
ment, for when the young hatched January 15th, 1884, nearly all were seen to be suffer-
ing from what has been termed dropsy or blue swelling of the yelk-sac*. As the same
cross had been made December 24th, 1881, but with older parents, and that without

* Prou. Zool. Soc. 1884, p. 370.

WHICH HAVE NEVEE DESCENDED TO THE SEA. 465

occasioning dropsy, one is irresistibly led to the conclusion that hybridization had
nothing to do with these results, which may reasonably be attributed to the immaturity
of the male parrs*. As a further confirmation of this view, parrs in 1881 taken from the
same lot were found prolific, but with comparatively few cases of dropsy occurring.
Among these 1000 dropsical fish, only about 100 lived out the year.

Thus the great mortality from eggs of young mothers was during the incubating stage,
hut from young males one season older it seems to have shown itself among the fry;
passing over another year, we do not lind so great a mortality among the eggs nor
disease of the young. Bearing on this I may allude to the ova of American charrs
having been milted from a Scotch charr in November 1882 ; one of the progeny gave 140
eggs on November 12th, 1881, or a little under 2 years of age. Only six feeble little
ones hatched on February 3rd, 1885. At the same time milt was obtained from some of
these hybrids and used to impregnate 4500 eggs of Lochleven Trout ; these hatched on
February 2nd, but the mortality was large, being upwards of half the eggs, while there
were many deformities and some dropsies among the offspring. Males are evidently
more matured for breeding-purposes than are females of the same age. I would here
advert to a remark I published in 1883 t, that in the fresh waters of India due to
indiscriminate net-fishing the young (of many forms of fish) have to be raised from ova
of such as are merely one or two seasons old, while the younger the parent the smaller
the eggs, and this is probably one mode in which races of fish deteriorate.

We now arrive at the question, Why is it that Salmon reared in ponds of fresh water
or where they are unable to descend to the sea are variously spoken of? Yarrell tells us of
Salmon fry being turned into a pond in 1830, and in 1833 several were taken 2 lbs. to 3 lbs.
in weight and perfectly shaped, but he seemed to be in doubt whether they could
mature their roe in such localities. Easch thought that fish thus imprisoned, if the
lake were very large, would attain about the same size as the Salmon in the sea.
Dr. Murie has shown that an arrest of growth occurred in an aquarium at the Zoological
Gardens among young Salmon kept in restricted areas.

I have had the opportunity of examining Salmouoids kept in artificial pieces of water
of different extents for various periods of time, the results of some of which I propose
alluding to as bearing on this subject.

In October 1879 I received from Mr. J. Carrington, F.L.S., of the Westminster
Aquarium, an American charr (Salmo fontinalis) 9 inches long, which had been hatched
from eggs sent from Lake Iluron, and which had been reared by the late Mr. Frank
Buckland in his tanks at the Horticultural Gardens, South Kensington. This fish of

* As corroborative of the above view that the milt of these young Salmon-parrs was deficient in marital powers,
I may mention that on the same date 1000 eggs of the common Brook-Trout were milted from one of these Salmon
parrs which had been dead a few hours. But not a single egg fructified ; only 3 turned white in December, 3 in
January, and 15 in February, or a total of 21. On March 12th the remainder were still quite clear, but destitute of
any sign of a contained embryo.

t ' Indian Fish and Fishing.' Great International Fisheries Exhibition 1883 series, p. 27.

466 ME. F. DAY ON THE BREEDING OF SALMON FROM PARENTS

several (? 6) years of age bad a head of a longer proportion * than is seen in these fishes
kept under normal conditions. In short the fish was not well grown — there had been an
arrest of growth due to its various surroundings. At the same time I received from
Mr. F. Buckland four examples of undoubted Salmo salar, reared by him from eggs
received from Huningen, and which were collected from Salmon captured fortius purpose
from below the falls of Schaffhausenf . As year after year passed by and these fish were
still retained in the comparatively small amount of fresh water which sufficed to fill the
tanks in the Horticultural Gardens, they became lanky and apparently half-starved fishes,
similar to those which have been descrihed by Dr. Murie (I. c, and in his 2nd paper
P. Z. S. 1870, pp. 30-50, " Irregularity in the Growth of Salmon").

Can Salmon be reared in small pieces of water or ponds through which a good current
of water is kept flowing and wherein there is a sufficiency and variety of suitable food?
To this we have a reply from Howietoun, to which I shall presently allude, that Salmon
can be so kept and reared, but the size to which they will yet attain time only
can demonstrate.

Certain hybrid Salmon have been reared and kept for some years in the Howietoun
ponds: so far most have proved to be sterile, and the size to which they have attained
lias been unsatisfactory ; but the fish have all been well formed and not shown that kelt-
like appearance usually deemed so characteristic of Salmon retained in fresh water.
The experiments have been as follows : —

On November 25th, 1879, some eggs from the Lochleven Trout were fertilized from the
milt of the Salmon. Although many hatched, up to this period none have been found
fertile; the largest I have seen was on November 11th, 1881, and it was lGi inches in
length. They were kept in the island pond (which is 210x90 feet and from 10 to 11 feet
deep in the centre). On November 28th, 1883, these fish were removed to pond No. 11,
where they now remain. On December 21th, 1881, some Lochleven-Trout eggs were
similarly fertilized with Salmon milt, and they hatched March 9th, 1882. The young
fish were kept in a planked pond, 20 feet long by 5 feet wide, until March 13th, 1881, when
they were removed to the octagon pond at Craigend %, and the largest taken has been
10 inches long.

It seems to me that at Howietoun those hybrid Salmon raised under the same con-
ditions as some Salmo salar do not augment in size so rapidly as these latter fish ; thus
among the hybrids between the Lochleven Trout and Salmon hatched in March 1S82, in
November 18S3 the largest was a little over 7 inches long, while the young Salmo salar
in the next pond, a year older, were from 4 to 13| inches long. If future observations
shall give the same results it will be necessary to ascertain whether this arrest of growth
is due to hybridization (which is unlikely), paucity of space or of water supply, character
of the ponds, or has any connection with the food.

* Day, British and Irish Fish. ii. p. 110, pi. cxx. fig. 2.
t Journ. Linn. Soc, Zoology, xvi. 1SS2, p. 397.

t These fish hegan to show symptoms of commencing fertility in May 1SS5 ; and on July 4th, owing to the stream
giving out due to the hot weather, all but one died.

WHICH HAVE NEVER DESCENDED TO THE SEA. 467

I may now finally remark upon the facts that pure Salmon have been raised from eggs in
the ITowietoun fishery, that the young have grown to be parr, smolts, and grilse, and that
these have given eggs and the eggs so obtained have been successfully hatched. Many
sportsmen and fishermen are now reopening the old question that grilse are not young
Salmon *.

The general bearing of the preceding authorities quoted has been in favour of the evi-
dence that Salmon confined in fresh water, and especially within that of limited areas, were
incapable of reproducing their kind. On the other hand, the fact obtains that land-locked
Salmon, both in Europe and America, certainly reproduce their species in fresh water.

As touching the question at issue, the Howietoun experiments, then, have shown the
possibility of not only rearing parr and smolts, but also obtaining ripe ova and milt from
them, without their ever having gone to the sea. Thus it is demonstrated that descent
to the sea is not, as has been asserted, a dire physiological necessity.

I would here further observe that although time will be necessary before a definite
reply can be given as to how these young Salmon raised from the eggs of parents which
have not gone to the sea will thrive and how large such fish will become in freshwater
ponds, still any unbiassed investigator must admit that, so far as they have gone, the
experiments made at Howietoun among the Salmonidre are pretty conclusive on the
following points : —

(1) That male parrs and smolts may afford milt competent to fertilize ova, but when
from fish of the second season, or up to 32 months old, it is (? always) of insufficient
strength for strong and vigorous fry to be raised.

(2) That female smolts or grilse may give eggs at 32 months of age, but those which
are a season older are better capable of producing vigorous fry ; while for the purpose of
developing ova a visit to the sea is not a physiological necessity.

(3) That young male Salmonidae are more matured for breeding-purposes than are
young females of the same age.

(4) That although females under 24 months of age may give ova, such are of
little use for breeding-purposes, the embryos not becoming well developed or vigorous,
while the young when hatched are frequently malformed.

(5) That the size of the eggs of Salmonida? varies with the age and condition of the
parent; but, as a rule, older fish give larger ova than do younger and smaller ones.

(6) That among the produce of every female fish there may be found variations in the
size of the eggs.

(7) That from larger ova finer and more rapidly growing fry are produced : conse-
quently that, by selection of breeders, races may be improved ; while it is only where
segregation is well carried out that such selection is possible.

* On the table, exhibited at the meeting when this paper was read, was a specimen reared at Howietoun from a
Salmon egg, the identity of which with grilse I think was beyond dispute. There were also a parr and a smolt
shown from the same hatching, evincing the uutenableness of the assumption that the latter have an additional coat
of scales to what obtains in the parr.

SECOND SERIES. — ZOOLOGY, VOL. II. 70

468 MR. P. DAY ON THE BREEDING OF SALMON.

DESCRIPTION OF THE PLATES.
Plate LIII.

Fig. 1. Salmo salar in the parr livery, a male with ripe milt, hatched March 1881, killed November 27th,

1883. Drawn of natural size.

Fig. 2. Caudal fin shown in two conditions of expansion ; natural size.

Fig. 3. Salmo salar in the smolt livery, a female with eggs only just visible, hatched March 1881, killed

November 1883 ; natural size.
Fig. 4. Caudal fin ; natural size.

Plate LIV.

Fig. 1. Salmo salar, a grilse that had been just spawned, hatched March 1881, killed December 9th,

1884. Sketched from the specimen, and of natural size.
Fig. 2. Caudal fin ; natural size.

Fig. 3. Scale from above lateral line and below adipose dorsal fin, magnified 20 diameters. From a
photograph of the specimen on a glass slide

r 469 ]

XIX. Golfingia MacIntoshii, a new Sipunculid from the Coast of Scotland. By
E. Ray Lankester, M.A., LL.D., F.B.S., Jodrell Professor of Zoology in
University College, London, Fellow of Exeter College, Oxford.

(Plates LV. & LVI.)

Read 18th June, 1885.

V\ HEN I was staving at St. Andrews last summer (1884) my friend Professor
Macintosh, knowing my interest in the class to which it belonged, very kindly presented
me with an exceedingly remarkable Gephyrsean which he had obtained ten years pre-
viously from a friend who had dredged it in St. Andrews Bay, south of Montrose, at a
depth of ten fathoms.

The specimen was obviously something new, and was noteworthy, not only on account
of its external structural features, but on account of its large size.

The anticipations, based on its external appearance, were justified by a dissection of
the specimen, which I carried out in the intervals of exercise with the club and ball
sacred to the classic "green" of St. Andrews; and I have accordingly ventured to dedi-
cate the new genus of Sipunculid worms indicated by this specimen to the local
goddess whose cult is historically associated with the most ancient of Scottish seats of
learning. " Golfingia " forms an appropriate parallel to the Scandinavian genus of
Echiurid Gephyraeans called " Hamingia" by Koren and Danielssen.

External Features. — Golfingia, as exhibited in the spirit-preserved specimen before
dissection, presented the appearance drawn in PL LV. fig. 1. It measured five inches in
length, and consisted of a soft-walled cylinder of white silky surface, marked with dark
dots, as in Sipunculus punctatus. At either end of this soft-walled cylinder a hard dark
brown-coloured spout or smaller cylinder is observed. One of these is the "posterior
sclerite " or " scleropyge," the other is the " anterior sclerite " or " sclerorhynchus." In
the living state the proboscis or "introvert" which Golfingia possesses, like all other
Sipunculids, would issue from the sclerorhynchus in the way shown in fig. 2. Both
scleropyge and sclerorhynchus are modifications of peculiar structures occurring in
Aspidosiphon, and are found in the cylindrical form only in Golfingia. The sclero-
pyge is the same organ as the posterior " shield " of Aspidosiphon, whilst the sclerorhyn-
chus is represented by the anterior shield of Aspidosiphon and by the calcareous ring of
Cloeosiphon.

The scleropyge is shown in figs. 3, 4, 5, so as to exhibit its surface-markings. Its walls
are very thick and quite inflexible. It probably was moved as though hinged to the soft
body, and was used in burrowing in sand. The body-cavity is continued into it, and the
nerve-cord extends more than halfway along it, giving off numerous nerve-filaments
(PL LVI. fig. 11). It is imperforate. The sclerorhynchus is similarly shown in figs. 0, 7, 8.

SECOND SERIES. — ZOOLOGY, VOL. II. 71

470 PROF. E. EAT LANKESTEE ON GOLFINGIA MACINTOSHII.

In the median dorsal line, at the base of the sclerorhynchus, is placed the triangular anal
aperture (a). Symmetrically on each side are the external apertures of the paired nephridia
(brown-tubes), fig. 6, dl, dr. The sclerorhynchus is as solid and thick-walled as the
scleropyge. Its surface is marked by elongated quadrangular arese differing from the
rougher transverse rugce of the scleropyge. The anterior extremity of the sclerorhynchus
presents an opening (fig. 7). This is the orifice of invagination of the proboscis or intro-
vert, which, when withdrawn, carries with it the tentacular crown and mouth. The form
of these parts was made out by dissecting them in their introverted condition (see figs. 9,
10, 12).

The introvert is soft-walled, its outer surface being beset with numerous papillae, and
near the mouth with chitinized booklets (as in many Sipunculids). These are shown
from different regions of the introvert in figs. 17, 18, 19, 20, and 21. The introvert
measured, when extended, two inches and a half in length (fig. 2 d). At its free extre-
mity is placed the mouth e, surrounded by a circlet of six symmetrical tentacles. Each
tentacle is pinnate, carrying two rows of small processes (about twenty-four in all). A
single tentacle is shown in fig. 13 and part of one in fig. 14.

Internal Features. — The body-cavity (ccelom) is spacious, as in Slpunculus. The
characters of the coagulated ccelomic fluid could not be ascertained.

The long-muscles of the body-wall form a smooth continuous coat, i. e. they are not
divided into groups, as in Slpunculus, Phymosoma, and some species of Aspldosiphon.

The intestinal coils did not extend further than two thirds of the length of the body.
This may have been due to the action of the spirit ; but I am inclined to think not.
Thus, the posterior third of the body and the scleropyge were simply filled with ccelomic
fluid, and contained no "floating" viscera. The intestinal coils were held together and
to the body-wall by a delicate mesentery, as in Slpunculus. The position of the retractor
muscles of the introvert, and the relations of the brown tubes (nephridia), rectum,
oesophagus, and nerve-cord to these and to one another are shown by the drawings of the
dissections made (figs. 9, 10).

There are four retractor-muscles of the introvert, as in Slpunculus (not two only, as in
Aspidoslphon). Two of these are attached anteriorly and dorsally (h), and two posteriorly
and latero-ventrally (i).

The anterior pair arise on each side of the rectum, but not quite symmetrically. The
posterior pair arise over (and are perforated at their bases by) the nerve-cord, which lies
here on the animal's left side near the middle ventral line.

The anus of Golfingia lies at the base of the sclerorhynchus in the middle dorsal line,
and the ventral line corresponding to this is the ventral median line. The nerve-cord
does not occupy this middle ventral line, and the only internal structures which exhibit a
strict bilateral symmetry in their position are the nephridia (see diagrams, figs. 15, 16, g).
The nephridia lie right and left of the symmetrical sclerorhynchus, equally distant on
either side from the anus. In the dissection they are seen as shrivelled brown sacs (figs.
9 &11). I could not discover their internal openings. In the dissection (fig. 9) the four
retractor- muscles are seen attached to the introvert at the point I, and the nerve-cord (ff)

PKOF. E. EAT LANKESTEE ON GOLFINGIA MACINTOSH!!. 471

is seen attached loosely to the whole length of introverted proboscis {y). From the middle
point between the retractors passes backwards the oesophagus (I) lying in close proximity
to the rectum ; a little way back they are twisted round one another. A curious muscular
band (the spindle-muscle of Selenka) accompanies the rectum (figs. 9, 10, m), and is
attached, after running a length of two inches, to the wall of the intestine. A similar
muscular band exists in Sipunculus nudus (and most Sipunculids), arising in that animal
from the curious little csecal pouch which exists on the rectum at the distance of an inch
or so from the anus.

I could not find any trace in Golfingia of this caecum, nor of the " bush-like organs "
near the anus, which appear in Sipunculus nudus, possibly to represent in a rudimentary
form the posterior nephridia (cloaca! trees) of the Echiuridean Gephyrsea.

Owing to the imperfect state of preservation of the specimen, I am unable to state
whether any vascular system of one kind or another exists in Golfingia, nor am I able to
give any indication as to the position of the gonads.

The distinctness of Golfingia from all other genera of Sipunculid Gephyraeans will be
obvious at once to those who have studied this group of animals. The species I dedicate
to my friend Prof. Macintosh, not only in recognition of his vast services to marine
zoology, but in memory of a long-standing friendship marked by many acts of kindness
on his part.

The relation of Golfingia to other genera of Sipunculidean Gephyrseans will be best
appreciated by the reader if I make use of the synoptic table recently published by
Selenka, and assign Golfingia its place in a modification of that table. It is not far
removed from Aspidosiphon, but differs from that genus both in the form of its sclerites
and in the disposition of the retractor-muscles and the character of the oral tentacles.

SlPUNCULOIDEA.
(Slightly modified from " Die Sipunculiden " of Dr. E. Selenka, Wiesbaden, 1883.)
Marine worms with cylindrical elongate bodies, with obliterated segmentation ; only
in the larval stage sometimes lateral, serial pairs of bristle bundles are present. Larva
with a well-marked prostomial area, and a ciliate band, which is post-oral. The mouth,
lying at the fore-end of the body, is provided with tentacles into which the body-cavity
never penetrates, but only the vascular system ; in rare cases there are no tentacles.
Skin with numerous glands ; overlying the cerebral ganglion is often a pair of ciliated
tubercles. The long, tubular, alimentary canal is almost in all cases spirally rolled;
very seldom is it simply thrown into loops ; the anus is dorsal in position and far for-
ward. The front part of the body (introvert) is introversible by the action of retractor-
muscles ; anteriorly on the introvert are very generally present chitinous hooks. The
ccelomic fluid contains corpuscles. The closed vascular system encloses a corpusculated
fluid, and consists of one or two contractde sacs (hearts) — which accompany the
oesophagus and end blindly after a short course — and of the vascular ring and tentacular
vessels. The circulation of the blood is effected both by the contraction of the heart's
walls and by the action of ciliated cells ; in very exceptional cases there is no vascular

472 PKOF. E. EAT LANKESTEE ON GOLFINGIA MACINTOSHII.

system. There are no special respiratory organs. A pair of nephridial sacs (rarely
only one) are always present and always (?) have an internal aperture. The reproductive
organs lie in the form of transverse ridges at the base of the ventrally attached pair
of retractor-muscles (of the introvert) ; the eggs have numerous pore-canals in their
chorion ; sexes distinct. Free or living in shells or tubes. In all seas.

Key to the Genera of Sipunculoidea.

I. The longitudinal musculature of the body-wall is continuous and not divided into
separate bands. The retractors of the introvert vary in number from one to four.

1. Phascolosoma. — Two nephridia. Numerous tentacles surrounding the mouth in a circlet. A

complete intestinal coil, unattached posteriorly (only in .1/. Hanseni attached) ; a spindle muscle
is usually present ; only on the anterior intestinal convolutions are there one or more ligaments.
Adhesive papillae are always absent. Hooklets arc generally not present on the introvert. The
retractors are four, two or only one. Eggs spherical. In all seas.

2. Dendrostoma. — Two freely suspended nephridia. Only four to six pinnate tentacles. A complete,

posteriorly unattached, intestinal coil ; spindle-muscle always present; ligaments present only on
the anterior intestinal coils. Hooklets are present, but may be shed early in life. Four or two
retractors. The contractile sac (of the vascular system) generally has caeca upon it. Tropical
forms.

3. Phascolion. — A single ncphridium, only that of the right side, -which is attached to the body-wall

throughout its length. The intestine forms no spiral or only an incomplete one ; there is no spindle-
muscle, whilst the loose intestinal loops are attached throughout their length to the body-wall by
numerous ligaments. Adhesive papilla; arc often present. The retractors are not more than two.
Eggs spherical. Living in Gastropod shells or in tubes. In all seas.

II. The longitudinal musculature of the body-wall is divided into 17-41 separate bands.
Four retractores introversi.

1. 1'hvmosoma. — Body covered with papillae. Numerous filamentous tentacles, which seldom (or
never ?) surround the oral aperture, but are arranged away from the latter, dorsally in a three-
quarter-circle which is open dorsal-wards. No caecurn on the rectum. Hooklets nearly always
present on the introvert. Four retractors (in P. Ruppellii two?). Contractile sacs almost always
devoid of villi. Eye-spots are always present. Eggs elliptic, flattened, reddish. Small tropical
species.

5. Sipunculus. — Body without papillae. Tentacles surrounding the mouth in a circlet. Always one

or several caeca on the rectum (except in »S. edulis ?). Hooklets absent, only present in S. australis.
Eggs spherical. The individual tentacles thick, with an internal vascular network (not three
longitudinal vessels merely) ; generally two contractile sacs. Mostly large forms. In all seas.

III. In front of the anus and also at the hinder end of the body a distinct shield
{corneous thickening of the integument), or in front of the aims a calcareous ring or a
tube-like coruification, forming the base of the introvert, and another at the tail-end.
Hooklets sometimes present. Longitudinal muscles continuous or divided into bands.

6. Aspidosiphon.* — An anal and a caudal shield. Introvert eccentric arising from beneath the anal

shield ventrally. Tentacles small and not numerous, placed in a semicircle above the mouth.

* See Dr. C. Ph. Sluiter, " Beitriige zu der Kennfcniss der Gephyreen aus dem Slulayischen Archipcl.,'" Natuurk.
Tidsch. v. Nederlandsch-Indie, Deel xliii. (1884), pp. 20-88, with 4 plates.

PEOF. E. RAY LANKESTER OX GOLFINGIA MACINTOSH!!.

473

Intestinal spire traversed by a spindle-muscle, which is fastened at the hinder end of the body.
Only two retractors, ventral in position and often fused witb one another.

7. Cloeosiphon. — Directly in front of the anus a calcareous ring, from the centre of which the introvert

is extruded. Longitudinal musculature continuous. Hooklets bifid. Tropical.

8. Golfingia. — A praeanal corneous tube (sclerorhynchus) and a caudal corneous tube (scleropyge) .

The introvert issues centrally from the praeanal tube, which is, in fact, its basal portion indurated.
Tentacles sis, pinnate, surrouuding the mouth in a circle. Retractors four, two ventral longer, and
two dorsal shorter, the two pairs arising at some distance from one another on the body-wall.
Chitinized hooklets are present at the oral extremity of the introvert ; further back on the same are
found cylindrical and subspherical chitinized papillae. Longitudinal musculature continuous.
Intestine not coiled throughout in a spire, not fastened posteriorly ; a spindle-muscle present. Two
freely suspended nephridia. Vascular system uncertain. Known species large (eight inches long
when expanded).

IV. Only two lamelliform tentacles. Four retractors. Few intestinal loops, quite free.
No vascular system.

9. Petalostoma.

V. No tentacles. No vascular system. A single retractor,

10. Oxchn'eso.ma. — Introvert long. Body small, pear-shaped.

11. Tylosoma. — No introvert (?). Body cylindrical, thickly beset with papillae, which are more closely

set and larger at the front and hind ends of the body.

A single nephridiwm.

DESCRIPTION OF THE PLATES.

Letters apply to all figures.

a. Anus, in the median dorsal line.

b. The sclerorhynchus (anterior sclerite) .

c. The scleropyge (posterior sclerite).

, * > Right and left nephridial aperture.

e. Mouth.
/. Nerve-cord.

ff. Nerve-cord attached to the introvert.
g. Nephridia (brown tubes, oviducts or sperm-
ducts.)
h. Anterior retractors of the introvert.
i. Posterior retractors of the introvert.
k. Rectum.
/. (Esophagus.
m. Muscular band accompanying the rectum.

Musculature of the body-wall (smooth).

Epidermis and cutis.

Muscular attachment of the nerve-cord within

the scleropyge.
Six terminal filaments of the nerve-cord (cauda

equina).
The peri-oral tentacles.
Papillary surface of the introvert.
Point of attachment of the retractor-muscles to

the introvert.
Chitinized ridges on the cuticle of the introvert.
Hooks on the same.
Soft papillae.
Chitinized papillae.
The introvert.

Plate LV.

Fig. 1. The specimen of Golfingia Maclntoshii of natural size and appearance, after ten years' preserva-
tion in spirit.

2. Diagram of Golfingia Maclntoshii, with expanded introvert and tentacular crown : natural size.

3. Ventral view of the scleropyge.

4. Dorsal view of the scleropyge.

SECOND SERIES. — ZOOLOGY, VOL. II. 72

474 PEOF. E. RAY LANKESTER ON GOLFINGIA MACINTOSHII.

Fig. 5. Lateral view of the scleropyge.

6. Dorsal view of the sclcrorhynclms and anterior part of the soft body. Showing anus a and
position of nephridial aperture dl, dr.

7. View of the anterior aperture of the sclcrorhynclms.

8. Ventral view of the sclerorhynchus.

9. Dissection of the specimen of Golfinr/ia Maclntosldi, showing the position of the introvert y
and its retractor-muscles hh, the nephridia ggs the nerve-cord f, and the rectum k, and
oesophagus /. The body-wall has simply been cut along a line nearly corresponding to the
median ventral, and reflected on cither side.

10. The introvert and oesophagus have now been thrown forward, after cutting the latter and the
retractor-muscles so as to show the position of the anus, and the attachment of the nephridia
to the base of the sclerorhynchus internally.

Plate LVI.

Fig. 11. The scleropyge opened by a dorsal, median, longitudinal slit, so as to show the nerve-cord lying
within its cavil \ .

12. The introvert opened by a longitudinal incision, so as to expose its papillary surface s (the true
outer surface here seen as introverted) ; the tentacles ?• and the mouth e.

13. A single tentacle enlarged to show the pinnae.

14. Apex of a tentacle.

15. 16. Diagrams of transverse sections taken near the base of the sclerorhynchus, to show the asym-

metry of the arrangement of organs ; fig. 15 includes the introvert; fig. 1G is just behind it.

17. A portion of the integument from near the oral extremity of the introvert, viewed from the
outer surface and showing the mesh-like arrangement of chitinized ridges u, and numerous
booklets v, also the soft papillae w placed in the spaces of the mesh-work. It is a question as
to whether the soft papillae are an early condition of the chitinized hooklets, or are entirely
distinct structures. The position of the booklets on the ridges, and of the papillae in the space
enclosed by the ridges, favours the latter supposition.

18. A small piece of the same, showing clearly the relation of the hooklets to the ridges.

19. A similar portion of integument taken a little further back (the hooklets do not extend for more
than a quarter of an inch behind the oral aperture), showing the ridges u, more powerfully
developed, and no hooklets.

20. A similar portion of integiiment a little further back than fig. 19 (about three quarters of an
inch from the mouth), showing the chitinization of the papillae x, and the reduction of the
ridges, to pigmented tracts. The chitinized papillae are apparently the same structures as the
soft papillse w, of figs. 17, 18, 19.

21. Lateral views of three chitinized papdlaa from the same region.

[ 475 ]

XX. On tke Variations in tke Form of tke Cirri in certain Cvnatul*. ^™^
Carpenter, B.Sc, F.B.S., Assistant Master at Eton College. {Communicated j
W. Percy Sladen, Sec. Linn. Soc.)

(Plate LVII.)

Read December 3rd, 1885.

in the same individual. ,./)«/»<•*« with about

„ shown - V^Al^S^ Otl^i,^,

,. it. e ti!o mntni'p eirrus and no trace ot an opuubiua o\>

proportions from those of the matin e centro-dorsal of any

penultimate. Both these types are usually to be found on he c
Urably mature ***■ ^ ;; while they may a so ; ccui ^^^ penta.
which are borne by the Pentacrinoid larva ^^^ Common On the other band,
crinoids of other species the small mature for* » tte most commo ^^

when studying Actmometra parmcirra I could find no youn corn

type at all, though the ^^^^^^Tt^een the rudimental and
Now in Antedon pAaUmgvm not only is this difleren ^

the small mature cirri carried to a very much greater xte nl than
the sbape and number of the ^ointsv.y ^^^ in app_

ance(Pl.LVHhSs.7^^)t ,£»*«*»; whereas in reality tbey are

r:er^^ — — • As i have a large

n 4. i„ T Or, the Suedes of ^eZecrmws and #wcZwcnnws," Journ. Linn.

' « Descriptions of new or little-known Comatuke. I. On the fcpecies

Trans. Linn. Soc. 2nd ser. Zool. vol. ii. 1879, p. 60. ^

SECOND SERIES.— ZOOLOGY, VOL. II.

476 DR. P. H. CARPENTER ON THE CIRRI OE CERTAIN COMATHL/E.

series of examples of this species, formerly so little known, from ten different localities, I
have devoted a considerable amount of time and trouble to working out in detail the
variations of its cirri, both local and individual, and the results of this investigation are
explained in the following pages.

Antedon phalangium was described by Miiller : as long ago as 1841, though for a long
time but little was known about it. The original specimens which Miiller described
had been obtained at Nice and at Naples; but for many years afterwards no further
mention was made of its having been met with, either at these localities or anywhere
else. It was obtained off the Tunis coast by the ' Porcupine' Expedition of 1870, though
the fact was not recorded ai the time ; and it was not till 1875 that much attention was
directed to it. Professor Marion dredged it in the harbour of Marseilles, and gave a
careful analysis of its peculiarities, accompanied by some excellent figures.2

These have enabled me to identify it with the type from the Uoss-shire coast which has
been so long known to British naturalists by the name of Antedon celtica '■' ; and it is of
the differences which are presented by the cirri of the Scotch and of the Mediterranean
types respectively that I am about to speak.

Besides obtaining Antedon phalangium in great quantities at two localities on the
Tunis coast, the 'Porcupine' also dredged it at Carthagena, off Cape Sagres on the
Atlantic coast of Spain, and also off Mondego on the coast of Portugal. S :veral speci-
mens were likewise obtained by the S. S. ' Dacia ' on the Seine bank at a depth of 88
fathoms, and have been put into my hands by Mr. John Murray; while, thanks to the
kindness of Professor Marion and of Mr. W. Percy Sladen, I have further been able to
examine specimens from Marseilles and from Naples respectively.

Midler's description4 of the cirri in the .Mediterranean variety of this type ran as
follows: — " Die Ranken sind zur Grosse des Thiers ganz ausserordentlich lang mit 45
langen dunnen Gliedern. Endglied gestreckt ohne Dornchen der Innenseite ; die
Glieder, mit Ausnahme der ersten (an der Basis), sind 2-2Jmal so lang als breit." This
is very well shown in the excellent figure given by Marion 5 of some specimens from
Marseilles ; and he mentions that the cirri vary in length from 25 to 58 mm. while the
number of joints varies from 37 to 51. In Mr. Sladen's Naples specimen the longest
cirrus reaches (31 mm. and consists of 48 joints, the three longest of which measure 5 mm. ;
and it may be noted in passing that only one young cirrus remains attached to the centro-
dorsal of this individual. It is a trifle over 5 mm. long and belongs to the rudimental
type, consisting of a large number of small annular joints without auy trace of a terminal
claw. The longest cirrus which I have found in specimens from the Tunis coast measures
52 mm. and contains 17 joints (PI. LVil. tig. 7) ; while in one from the Minch there are 51

: " TTeber die Gattuug Comatula, Lam., und ihre Arten." Abkandl. d. k. Akarl. d. Wis.s. Berlin, a. d. J. 1847
(1849), p. 253.

2 " Draguages au Large de Marseille," Ann. des Sc. nat. 6' serie, /•».!. t. 8, L879, pp. 4n-l."> (of separate copy),
pi. 18.

"Note on the European ComatulsB," Zool. Anzeiger, 1881, Jahrg. iv. p. 520, and Proc. Roy. Soc. Edin. veil. Nii.
pp. 361, 362.

1 Loc i-it. p. i!5;i.

~° Loc.cit. pi. IS, figs. 11, lie, 11-/.

DR. P. H. CAEPENTEE OX THE CIRRI OF CERTAIN COMATUL.E. 477

joints, though the length is only 47 nam. (PI. LV1I. fig. 29), and in the most extreme form
from this locality there are 48 joints in a length of hut 35 mm. (PI. LVII. fig. 33). In the
Scotch variety therefore, while the number of joints is but little altered, they are relatively
very much shorter than in the Mediterranean examples.

Let us now study the mode of development of these joints and the transitional forms
between one kind of cirrus and the other. Por the typical form of cirrus in the species
we must of course take the long-jointed ones which were described by Miiller in the
Mediterranean variety (PI. LVII. fig. 7).

Type A. Long-jointed. — The three basal joints are wider than long, but the length
gradually increases from the first to the seventh and then remains tolerably constant for
some distance, decreasing slowly towards the end of the cirrus. But the terminal joints
ai'e not laterally compressed and wider than long, or even square, as is the case in the
cirri of most Comatnlae ; for they remain much longer than wide. In fact the width
diminishes slightly in the last few joints so that the end of the cirrus has a somewhat
tapering appearance, and there is a long terminal claw, without, however, any trace
of an opposing spine on the penultimate joint. Most of the cirri are of this type in all
the Mediterranean specimens of Antedon phalangium that I have seen ; and it is also
the prevailing one in those from off Cape Sagres and off Mondego. But in those
dredged by the ' Dacia' on the Seine bank it is very largely intermixed with another type
of cirrus (PL LVII. fig. 23). the terminal joints of which are rather short and broad,
somewhat as in those from the Boss-shire variety (PI. LVII. fig. 33).

The ' Dacia ' dredgings, and in a less degree also those of the ' Porcupine,' yielded several
individuals in various stages of maturity; and I have therefore been able to study the
development of the cirri both in quite young individuals and also in those older ones in
which new cirri are continually being produced round the margin of the centro-dorsal.
So far as I have been able to make out, these typical long-jointed cirri are developed on
the rudimental plan. An early stage of their formation is shown in PL LVII. fig. 1. The
fifth and sixth joints are already distinctly longer than wide; but their successors
become rapidly shorter, and the rounded end of the cirrus is made up of a series of
annular joints without any trace of a terminal claw. If we make allowance for the meat
difference between the cirri of the two species, we may place this form between stages
C and D of the developmental series described by Dr. Carpenter in Antedon rosacea.
Its further development is shown in PL LVII. figs. 3 and 6. In tbose two cirri which
measure 14 and 22 mm. respectively, and especially in the latter, which has some fifty
joints, the lower joints present all the characters of maturity ; but the later ones are
mere discs, and it is only in the larger one (fig. G) that there is any indication of a
terminal claw and that but a very slight one. The claw becomes more marked in the
older cirri shown in figs. 4 and 5, and the later joints lengthen out till their shape more
nearly resembles that seen in the mature cirrus (PL LVII. fig. 7). As in the case of tin1
rudimental cirri of Actinometra parvicirra l the addition of new joints seems to take
place in the outer half of the cirrus, those at its base presenting most of the characters
of maturity at quite an early date.

1 Trans. Linu. Soe. 2nd ser. Zool. vol. ii. p. 60.

478 DR. P. H. CARPENTER ON THE CIRRI OF CERTAIN COMAIJL^.

It sometimes happens that the terminal claw makes its appearance when the cirri
are much shorter than that represented in PI. LVIL fig. 5. An instance of this is
shown in fig. 2, and the later stages are seen in figs. 8 and 9. The claw is at first
almost central in position (figs. 6, 9), hut it gradually moves over towards the dorsal side,
though no opposing spine is developed, as seen in PL LVIL figs. 5 and 8.

I have not found any cirri of this long, smooth, and very long-jointed Mediterranean
type among the Scotch individuals of Antedon phalangittm ; the nearest approach to it
being the cirrus, of which the end is shown in PL LVIL fig. 16. But this rather belongs
to the intermediate type (B) next to be described, more southern examples of which are
shown in figs. 15 and 17.

Type B. Intermediate. — Besides these greatly elongated cirri with numerous joints,
the terminal ones of which are very long and slender, most of the specimens from the
Mediterranean and from the Seine bank have somewhat shorter and smaller cirri with
fewer joints, the terminal ones of which are relatively broader than in the typical form.
Of this kind there are many varieties ; while there is often a more or less distinct
opposing spine on the penultimate, and sometimes also on the previous joint as well.
Fig. 17 represents a moderately young one without an opposing spine, and fig. 19
another still younger in which it is tolerably distinct. In both of these cirri the joints
exhibit all the characters of maturity, although they are less numerous than in the
longer but very undeveloped rudimental form seen in fig. 6. Older cirri of this pattern
may either have no opposing spine (tig. 15), though the dorsal edge is very slightly
serrate, which is not the case in the type (fig. 7) ; or it may be present, sometimes with
another behind it (fig. 11). In this specimen the joints forming the outer third of the
cirrus are distinctly longer than wide; but in those shown in figs. 10 and 11 this is less
evident, and in the former the opposing spine is double. This form with broader terminal
joints sometimes occurs in the Scotch variety of the species (PL LVIL fig. 16); and it
thus presents an approach to the cirri with still shorter joints which are so characteristic
of this local variety (PL LVIL fig. 33). I have not found sufficient material for working
out the development of this shorter but still long-jointed form of cirrus ; but I think
that figs. 12, 13, and 18 represent some of its earlier stages. The second of these, which
is not half the length of the very immature form shown in fig. 3, came off the smallest
individual of this species which I have seen.

Type C. Square-jointed. — Besides the more or less long-jointed cirri typical of the
Mediterranean variety, the specimens dredged by the ' Dacia ' on the Seine bank bore a
number of cirri much more like those characteristic of the Boss-shire form. The end of
one of these is shown in PL LVIL fig. 23. It has 57 joints which gradually decrease in
length from the middle of the cirrus onwards, and the short later ones are somewhat
compressed laterally, but about as wide as long. In this cirrus there are spines on both
the joints before the claw, while there is a third spine on the rather smaller cirrus shown
in fig. 24. On the other hand there may be only one spine (fig. 25), or none at all, as in
that shown in fig. 27, which also has a very greatly reduced terminal claw. But the
relative shortness of the joints in the outer third of the cirrus is very marked in all
these examples.

DR. P. H. CARPENTER ON THE CIRRI OF CERTAIN COMATUL.E. 479

This, which may be called the square-jointed type, furnishes the greater number of
the cirri in the Scotch specimens. A very characteristic example consisting of 51 joints,
but only 47 mm. long, is shown in fig. 29, and another variety with the terminal joints
longer in proportion to the width than those just below them in fig. 28; while fig. 21
shows another form with shorter joints and two opposing spines. The three youngest
cirri of this type that I have been able to find are shown in figs. 20, 22, and 2G. The
first two of these are from quite young individuals ; but the third is from a much older
one. The opposing spine and the carination of the later joints which is prominent in
the two former is less evident in the larger cirrus, which has rather the appearance of
having been developed on the rudi mental plan (compare fig. 31) ; while the two small
cirri of the young individuals clearly belong to the small mature type of Dr. Carpenter.
The smaller of the two is distinctly shorter, though with wider terminal joints than a
rudimental cirrus of type B, from another young individual at the same locality.

Type D. Short-jointed. — Besides the square-jointed form of cirrus just described, which
is the prevailing one in the Scotch variety of Antedon phalangium, several others may
be found with the later joints distinctly wider than long. The most extreme form of
this short-jointed type which I have seen is shown in fig. 33. Figs. 30-32 seem to
represent earlier stages of its development, though I am by no means clear that the last
one, if not the other two as well, may not belong to the square-jointed type shown in
fig. 29. They are interesting, however, as furnishing an example of the rudimental
process of development which we have also studied in the long-jointed Mediterra
variety (figs. 1-6) ; and a comparison of the two brings out very clearly the difference
between the latter and the short-jointed northern form in their immature condition.

Thus in a young cirrus from the Minch shown in fig. 31 there is a distinct terminal
claw, and the later joints have almost assumed their final shape; whereas in the much
longer but very undeveloped cirrus of the southern type (fig. 6) there is only the merest
trace of a terminal claw, and the later joints are still in a discoidal condition. In a later
stage, however (fig. 4), they have lengthened out and begun to look more like those of
the full-grown cirrus (fig. 7); while in the corresponding stages of the short-jointed
variety this great change of form has not taken place, thougli the size of the joints lias
considerably increased.

It is difficult to account for these striking variations in the form of the cirri of
Antedon phalangium. We have here to deal with something more than a i imor-

phism of the cirri in any given individual, such as I have noted in the boreal Comatulaa
Antedon hystrix, A.prolixa, and^. quadrata, &c. ; for the geographical distribution h
be considered as well. While the two intermediate types of cirrus occur in individuals
from every locality, northern and southern alike, the long-jointed form characteristic of the
Mediterranean is not to be found in any of the Scotch specimens; while those from the
Mediterranean, the Seine bank, and even from 220 fathoms in the Atlantic; off the
Portuguese coast, have none of the quite short-jointed cirri, such as thai shown in
fig. 33. Neither is it in any way a bathymetrical variation. The Scotch specimens
were obtained in the Minch at a depth of 60-80 fathoms. The ' Dacia dredged on the

SECOND SERIES. — ZOOLOGY, VOL. II.

480 DR. P. H. CARPENTER ON THE CIRRI OF CERTAIN COMATULuE.

Seine bank at 88 fathoms ; while the depth on the Tunis coast where the ' Porcupine '
dredged varied from 50 to 120 fathoms, and Marion found the species to be'most abundant
at Marseilles in 100-200 metres. Temperature variations again seem to have nothing
to do in causing this difference, for there arc plenty of long-jointed cirri in the
individuals brought up from 220 fathoms at a temperature of 52° Fahr., which is very-
near that observed in the Minch, while the Mediterranean water on the Tunis coast
was much warmer.

These variations in the form of the cirri are the more remarkable because in most
Crinoids these organs are very constant in their characters. In fact, among the
Pentacrinkbe there are several species of which the cirri are so similar that they can
hardly be differentiated when removed from the stem. In the genus Actinometra the
the cirri are, as a rule, lew in number ami poorly developed ; but the lower pinnules of
the arms often reach a considerable size and are provided with keels and processes of
different kinds ; and in some species, such as Actinometra Solaris and its allies, indi-
viduals from different localities present much the same sort of variation in the pinnule-
characters as we have seen to occur in the cirri of Antedon phalangium.

DESCRIPTION OF PLATE LVII.

[All the figures are x 5, except figs. 13, 20, 22, which are X 10.]

Type A. Long-jointed.

Figs. 1-6. Young cirri of the large rudimental type ; from the Seine bank.
Figs. 8, 9. Similar forms from the Tunis coast.
Fig. 7. Adult cirrus from the same locality.

Type B. Intermediate.

Figs. 10, 11, 13, 18. From the Seine bank. Fig. 13 represents a very young cirrus, probably

of the small mature type.
Figs. 12, 17, 19. From the Tunis coast.
Figs. 14, 15. From oft' Cape Sagres.
Fig. 16. From the Mincb.

Type C. Square-jointed.

Figs. 20, 22. Young cirri of tbe small mature type ; from the Seine bank.
Figs. 23-27. Older cirri from the same locality.
Figs. 21, 28, 29. Older cirri from the Minch.

Type D. Short jointed.

Figs. 30-32. Immature cirri of the large rudimental type; from the Minch.
Fig. 33. Adult cirrus from the same locality.

INDEX.

AcaDthidse, 249.
Aoaridse, 250.

Aeentropus, analogy of, 279.
Acherontia, (footnote) 400.
Acridium, larva of, 404.
Actias Luna, (footnote) 332.
Actinocrinus, 21, 439.
moniliformis, 21.
Actinometra, On the Genus, with a
Morphological account of a new
Species (A. polymorpha) from the
Philippine Islands, 1.
Basals of, 93-100.
Centrodorsal Plate of, 61-75.
Characters of, 17, 29.
Distichal and Palmar Series of,

44.
Dorsal Cirrhi of, 58^61.
Historical, 1.
Pentagonal hase of the Calyx,

75-93.
Eadials, 2nd, 3rd, 110-112.
Skeleton, with Ligaments and

Muscles, 53-58.
Specific diagnosis of type, 29.
Syzygia in arms of, 45-49.
Actinometra, 1, 2, 3, 5, 7, 10-21,
25, 26, 27, 29, 34-40, (footnote)
41, 53, 58, 61, 62, 65, 70, 71, 72,
74, 75, 77, (footnote) 78, 81, 82,
84-87, 94, 95, 96, 100, 101, 103-
108, 435, (footnote) 475, 480.
armata, 50.
Bennettii, 27, 29.
brachiolata, 27.
echinoptera, 27.
fimbriata, 27, 28.
imperialis, 11, 15.
meridionalis, 27.
multifida, 16, 22, 27, 28, 29,
45.

Actinometra

multiradiata, 13, 16, 21, 24,

26, 27, 28.
parvicirra, 27, 28, 475, 477.
pectinata, 27, 28, 42, 62, 67,
82, (footnote) 85, 88, 89, 90,
91, 92, 98, 99, 100, 102, 104.
polymorpha, 1, 17, 18, 22-29,
31, 34, 38-53, 56, 58-67,
70, 84, (footnote) 85, 88, 90-
93, 98-102, 104, 110, 111,
112.
protectus, 19.
purpurea, 27.

robusta, 27 (footnote), 29, 61,
62, 67, 70, 71, 83, 86, 87,
88-92, 98.
rosea, 27.

rotatoria, 27, 28, 45.
Solaris, 11, 12, 15, 18, 19, 20,
22, 23, 25, (footnote) 26, 27,
28, 38-42, 62, 66, 67, 83, 84,
88,91,92,98,100,104,480.
tenax, (footnote) 29, 45.
trachygaster, (footnote) 29.
trichoptera, 27.
Wahlberghii, 11, 12, 18, 19,

25, 26, 27
Actinometra?, 19, 25, 27, 29, 31, 35,
36, 42, 44, 45, 61, 62, (Pentagonal
base of Calyx of) 75, 88.
^Eshnia, 393 ; eyes of, 414, 417.

grandis, 4^0.
Agrion, 393, 401, 410, 414, 417, 419.

virgo, 419.
Alachtherium, 213, 214, 215.

Cretsii, 214.
Albania, 132.

formosa, 125, 128, 133, 140.
Alecto, 4, 5, 6, 9, 10, (footnote) 11.
12-20, 26, 27, 29.

SECOND SEEIES. — ZOOLOGY, VOL. II.

Alecto

alticeps, 13.
carinata, 4.
Eschrichtii, 13.
europrea, 4, 12.
horrida, 4.

multiradiata, 12, 13, 18, 19.
parvicirra, 51.
phalangium, 13.
Amblystoma, 187, (footnote) 207, 193.
opacum, 207.
punctatum, 208, 209.
"Weismanni, 208.
Amphibia, 206.

Amphistoma, 224, 251, 252, 253.
asperum, 241, 242.
Collinsii, var. Stanleyi, 238, 239.
Hawkesii, 224, 238, 239, 241.
ornatum, 224, 240.
papiUatum, 224, 240, 241, 242,

258.
pyriforme, 241.
Amphiuma, 188, 206, 208, 209.

tridactylum, 208.
Anaides lugubris, 20S.
Anaphe, Contributions to the know-
ledge of the Genus, by Lord Wal-
singham, 421.
Anaphe, 421, 422, 424, 425.
ambrizia, 422.
Carteri, 423, 426.
infracta, 422, 426.
panda, 422, 423, 424, 425, 426.
reticulata, 422, 425.
venata, 422.
Anatomy of Ants, on the, by Sir J.

Lubbock, 141.
Anatomy of Polynoina and Polynoe
(Lepidonotus, Leach) clava, of Mon-
tagu, by A. G. Bourne, 347.
Anergates atratulus, 151.

75

482

INDEX.

Anguis fragilis, (footnoto) 193.
Anoplura, 246.

Antedon, 2-5, 7, 10, (footnote) 11,
14, 16-20, (footnote) 23, 26, 29,
30, 31, 39, 41, 42, 50, 53, 56, 58,
61, 68, 69, 70, 74, 75, 76, (footnote)
78, 80, 81, S2, 94, 95, 100, 1U3,
104, 435, 440, 475.
adeonoe, 29.
armata. 29.
articulata, 29, 45.
bieolor, 21).
carinata, 29.
celtica, 29, 61, 63, 65, 70, 71,

77. 81, 82, 84, 85, 86, 88,

91, 92, 93, 95, :■>>, (footnote)

100, 1 1 '3. 476.
cubensis, 29.
dividua, 29.
Dubenii, 29.
Eschricbtii, 29, 34, 36, (footn ote)

70, 71, 82, (calyx) 100, 101,

104, 105, 475.
europrcus, (footnote) 74.
gorgonia, 4, 5, 16.
Hargenii, 29.
hystrix, 475, 479.
Jacquinoti, 2!'.
macrocnema, 29, 61, 71.
meridionalis, 20.
Milberti, 29.
Milleri, 29.

palmata, 23, 24, 29, 45.
petasus, 29.
phalangium, 29, 475, 470, 477,

478, 479, 480.
Philiberti, 29.
polyactinis, 29.
prolixa, 47"). 470.
protectus, 19.
quadrata, 475, 479.
Reynaudii, 29.
rosacea, 2, 5, 16, 20-25, 29,

32, 35, 40, 42, 5S-60, 70-

96, 100-106, 110-112, 475,

477.
rosaceus, 16, (footnote) 33, 34,

6S.
rubiginosa, 29.
Sarsii, 29, 75.
Savignii, 23, 24, 29.
semiglobosus, (footnote) 74.
(Comatula) rosacea, (footnote)

475.
Anthea lloylei, 332, (tegumen of)
332.

Antinoe, 353.

prasclara, (footnote) 353.
Wahlii, (footnote) 353.
Ants, anatomy of, by Sir John Lub-
bock, 143.

front legs and their muscles,

146.
mcsothorax and middle legs,

1 19.
muscles of head, 145.
organs of prothorax, 148.
posterior portion of thorax,

150.
prothorax, 143.
tibial organs, 147.
Anura, 171. 172, 183, 207.
Anns in Ledidoptera, site of, 276, 290,

304, 305, 336.
Apathus, 358.
Aphis, 412.

rosa;, 412.
Apbrodita clava, 317.
Apiocrinida', 14, 107, 10S, (footnote)
L09, I 1".
Lnites, •".
Apiocrinus, 57, 90, 110.
ellipticus, L08.
mespiliformis, 107.
Milleri, 107.

obconicus, (footnote) 23, 57, 90.
rosaceus, 1"7.
rotundus, 90.
Arachnida, 224, 234,
Arachnids, 365, 306, 414, 416.
Arctia, 425.

villiea, 425.
Arctiidae, 125.
Arctiomorpha, 122, 425.
euprepiaeformis, 422.
Argynnis Laodice, 334.
Arthropoda, 113, 147, 224,240,250,
366.

morphology of the eyes of,

414.
segregate and continuous retina

of, 409.
visual organs of, 410.
Arthropods, 398, 401, 405. 406,409,

412-414, 410,417.
Articulata, 5, 14.
Arvicola amphibius, 202.
Ascarides, 225.
Ascaris, 224.

lonchoptera, 224, 225, 253.
Aspidoeotylus, 241.
Aspidosiphon, 409, 470, 471, 472.

Asterias, 2, 3, 4.

multiradiata, 3, 4, 11, 12, 16.
pectinata, 2,3, 11, 12, 15.

Asteridae, 1. 4.

Asterites liberi, 5, 6.

Asteroidea, 2.

Astrophvton, 1.

Asymmel ry, in Hesperiadoe, (footnote)
298 ; in uncus of Papilio Machaon,
31 »3 : in valves of Papilio Schnieltzi,
2'.i-.

Atck'crinus, (footnote) 475.

Attacus Atlas, (footnote) 27S.

Aves, 25; i.

Axolotl, 189.

" Hack and front," terms how nsed,

273. 296, 305, 324,337.
Bathycrinus, 107, 109.
Bathyerges maritimus, 262.
Batrachia, L68, 173, 175, 187, 191,

193, 199, 207, 2o:i.
I ;.ii i ai bo eps, 209.

attenuates, 208.
Benham, W. 0. S., (in the Testis of

Limulus, 363.
Bombinator igneus, 207.
Bombus, 142, 358.
Bombyces, 42 1.
Bombyx mori, 42 I.
Bomella, 125, 138.
digitata, 139.

marmorata, 125, 138, 140.
Bot of Gastrophilus elephant is, 247.
Bourne, A. G., On certain Points in
the Anatomy of the Polynoina,
and on the Polynoe (Lepidonotus,
Leach) clava of Montagu, 347.
Brachymystax coregonoides, 462.
Branchippua, 402.

Breeding of Salmon from Parents
which have never descended to the
Sea, On the, by F. Day, 417.
Bridelia micrantha, 421.
Bufo, 172.

agua, 1S7.
paiitherinus, 209.
vulgaris, 172, 209.
Burmeister, on harpe, 273 ; on meco-
nium, 269, 288 ; on turgescence in
generative organs of Lepidoptera,
(footnote) 286 ; on valve, 205.

" Caducibranehiate " Urodele, 173.
Caducibranchs, 177, 183, ISO, 188,
190, 191, 19H, 202.

INDEX.

483

CaUidry as, (footnote) 271; prehensors,
332.

Eubule, 332, 333.
Statira, 333.
Caput-medusse, 2, 3, 4, 8.
asteria, 68.
brunnum, 2, 5.
cinereum, 2.
Carinatas, (footnote) 181, 193.
Carpenter, P. Herbert, On the Genus
Actinomctra, Mull., with a Mor-
phological Account of a new
species (A. polymorpha) from the
Philippine Islands, 1 .

On three new Species of Meta-
crinus, with a note on a new
Myzostoma, by Prof. L. von
Graff, 435.
On the Variations in the Form
of the Cirri in certain Coma-
tulse, 475.
CaBsidaria, 220.

Caterpillars, eyes of, 395, 404, 414.
Cavia aperaea, 260, 261.
Centetes, 261.

ecaudatus, 261.
semispinosus, 261.
Ceratodus, 206.
Cercopithecus callithricus, 260, 261,

264.
Cervus columbianus, 358.
Cestodes, 250.
Chaetopods, 445.
Characters of the Genus Actinometra,

17.
Chelocrinus, (footnote) 109.
Chromodoris, 124, 125, 127.
Alderi, 125, 132, 140.
aureo-purpurea, 125, 129, 130,

140.
Bulloekii, 125, 128, 140.
iris, 125, 127, 128, 140.
funerea, 125, 131, 140.
runcinata, 128.
Semperi, l-s.
tenuis, 125, 130, 131, 140.
Tryoni, 129, 130.
tumulifera, 125, 130, 140.
Cidaridae, 73.
Cirripides, 444.

Clasping-organs ancillary to Genera-
tion in certain Groups of the Lepi-
doptera, On the, by P. H. Gosse,

265.
Clbeosiphon, 469, 473.

Cnethocampa pityoeampa, 425.
Cobbold, Dr. T. Spencer, The Para-
sites of Elephants, 223.

On Simondsia paradoxa and on
its Probable Affinity with
Spkaerularia bombi, 357.
Cockroach, 401, 402, 405, 407, 420.
Ccelacanthus (C. Tingleyensis), from
the Yorkshire Cannel Coal, On a
new Species of, J. W. Davis, 4H7.
Ccelacanthus, 430, 432.
lepturus, 430, 431.
Tingleyensis, 433.
Colei, (footnote) 270.
Coleus, (footnote) 270.
Colias, (prehensors) 333, 334, 405.

Edusa, 334.
Collingwood, C, On some New Species
of Nudibranchiate Mollusca from
the Eastern Seas, 123.
Comaster, 7, 8, 9, 10, 12, 14, 16, 68,

73.
Comatoe St elite, 1.

Comatula, (footnote) 2, 3-10, 12-15,
19-24, 26-29, 32, 33, 35, 37,
(footnote) 41, (footnote) 51, 54, 55,
57, 62, 63, 67, 68, 71-75, 87, 90,
93, (basals of) 93, 104, 106, 107,
109, (footnote) 476.
adeonae, 5.
barbata, 7.
Bennett i, 10.
brachiolata, 5, 27.
brevicirra, 28, 51.
brevipinna, 29.
carinata, 4, 5.
Cumingii, 29.
echinoptera. 27, 28.
elongata, 29.
europsea, 12.
filiformis, 8.
fimbriate, 5, 7, 8.
flagellata, 29.
japonica, 29.

mediterranea, 2, 3, 5, 6, 7, 9,
10, 11, 13, (footnote) 33, 67,
68, (footnote) 71.
meridionalis, 20, 27, 28.
multifida, 9.
multiradiata, 5-10, 12, 13, 10.

18, 19, 21, 28, 68.
novae Guinea?, 29.
parvicirra, 28.
pectinata, 8, (footnote) 11.
pinnata, 8, 9.

Comatula

purpurea, 27.

rosacea, 4, 16, (footnote) 68, 93.

rosea, 27, 28.

rotalaria, 5, 8, 12.

simplex, 28, 51.

Solaris, 5, 11, 12, 15, 19.

tenella, 8.

tessellata, 29.

timorensis, 29.

trichoptera, 26.

"Wahlberghii, 11.

(Actinometra) Solaris, 12.

( ) Wahlberghii, 11.

(Alecto) europtea, 12.

( ) multiradiata, 12.

(Antedon) rosacea, 16, 93.
Comatute, 2-11, 16, 18-28, 34, 39,
43, 50, 62, (footnote) 85, (footnote)
100, 104, 125, 435, 439, (footnote)
475, 476, 477, 479.
Comatuhe, On the Variations in the
Form of the Cirri in certain, P. H.
Carpenter, 475.
Comatulidje, 1, 2, 14, 73.
Comatulinrc, 9, 13.
Compound Eye, Development of,

(Lowne), 411-414.
Compound Vision and the Morphology
of the Eye in Insects, by P>. T.
Lowne, 389.
Copilla, 416.
Corethra, 403, 412.

plumicoruis, 403, 420.
Cornea and Lens in Insect eye,

401.
Coryceidae, 416.
Corythaix, (footnote) 193.
Costata, 14.
Crinitte, 1, 4.
Crinoidea, 2, (footnote) 5, 8, 67.

(footnote) 475.
Crinoids, (TesseUate) 21, 435, 480.
Crocea zaffarana Neapolitanorum, 2.
Crustacea, 366, 402-405, 409. 411,

413, 414.
Cryptobranch stage, 170, 188.
Cryptobranchus japonicus, 208.
Cryptus formosus, (footnote) 423.
Crystalline Cone in Insect eye, 404.
Cupressocrinus, 75, 85.
Curtains in Omithoptera Remus, 336 ;

in Papilio Erithonius, 315.
CyathocrinidaB, 14.
Cyathocrinus, 439.

75*

484

INDEX.

Cynops, 180, 186, 1S8, 190, (foot-
note) 193, 199.

pyrogaster, 1S3, (skull) 183.

Davis, J. W., On a new Species of
Ccelacanthu8 (C. Tingleyensis)
from tho Yorkshire Cannel Coal,
427.
Day, F., On the Breeding of Salmon
from Parents which have never
descended to the Sea, 447.
Decacnemus, 2, 7, 8, 10, (footnote)
11,13, 14.
barbata, 2, 7.
fimbriate, 2, 3, 7.
rosacea, 3, 4.
(Antedon) rosacea, 3.
Decameros, 14.

Decempeda Cornubiensium, 2.
De Haan, on harpe in Lepidoptera,
273 ; on Papilio Polydorus, 330 ;
on Papilionidoe, 20(3 ; on scaphium
in Lepidoptera, 277.
Dendronotidce, 138.
Dendrostoma, 472.
Desmoguathus, (footnote) 175, 187,
(footnote) 193, 195, 204.
fuscus, (skull) 202, 207, 208.
Digastric Muscle, its Modifications
and Functions, On the, G. E.
Dobson, 259.
Dioptron, 391, 413, 414, 416, 418,
Anatomy and Functions of(Lownc),
392-407.

segment of, 391.
Dipnoi, (footnote) 193, 206.
Diptera. 400, 104, 405, 412, 414, 425.
Dipteron, Egg of, 249.
Distoma, 243, 244.

hepatieum, 242.
Distomum, 225.

hepatieum, 225, 242.
elephantis, 242.
Diurnal Flies, 407, 408.
Dobson, G. E., On the Digastric
Muscle, its Modifications and
Functions, 259.
Dochmius Sangeri, 224, (hood and

rays) 236, 258.
Dorididae, 126.
Doridopsida3, 134.
Doridopsis, 125, 134.

arborescens, 125, 134, 135, 140.

nigra, 135.

rubra, 125, 135, 140.

Doris, 124, 126.

aspersa, 131.

cerebralis, 127, 131.

coccinea, 124.

crescentica, 125, 126, 127, 140.

exanthema, 124, 127.

flammea, 124.

pecten, 125, 126, 139.

petechialis, L30.

smaragdina, 128.

sordida, 124.

tristis, 121.

tubcrculata, 1 24.
Doritis Apollina, 267.
Dragonfiies, 394, 400, 401, 405,

407, 409,411,413,417,419.
Dynastor, 334, (prchensors) 334.

Darius, 334.

Eaton, on Spermatozoa in Ephemera,

288.
Echinidse, 73.
Echinococcus, 367.
Echinoidea, 7:i.
Echinorhynchi, '■'•I':' >.
Echiurid Gephyrseans, 409, 471.
Ectozoa, 224, 249, 254.
Egg of Dipteron, 249.
Elephas iudicus, 225, 226, 230, 231,
233, 235, 236, 237, 239, 240, 242,
247.
Ellipsoglossa, 209.

nasvia, 208.
Emu, 198.
Encrinites, 5.

Encrinus moniliformis, (footnote) 23.
Entomostracous Crustacean, eye of,

413.
Entozoa, 224, 231, 253, 254, 256,

357.
Eolis, 124.
Ephemeridse, 288.
Epomophorus, 260.

comptus, 260,261.

franqueti, 260, 261, 264.

gambianus, 260.

labiatus, 260.

macrocephalus, 260.

minor, 260.

pusillus, 260, 261.
Equites, 205, 277.
Erinaceidoe, 261.
Erinaceus, 262.

europoeus, 260, 264.
Eucalyptocrinus, 21.

Eudiocrinus, (footnote) 475.

varians, 475.
Eugeniacrinidso, 14.
Eugeniaerinites, 5.
Eugeniacrinus, 14, 108, 109, 110.

caryophylhitus, 108.

Eagenowii, 108.
Euryades, 267, 329.
Euryale, 1, 4.
Eurycus, 267, 329.

Cressida, 267.
Extracrinus (Jurassic), 441.
Eye (Insect)

Cornea and Lens in, 401.

Crystalline Cone in, 404.

Great Rods in, 406.

Fasciola, 224, 24:'., 244.
gigantea, 243.
nepatica, 243, I'll, 250.
Jacksoni, 224, 242, 243, 244,
250, 252,254, 256, 258.
Filaria, 224, 358, 367-388.
immitis, 382.
medinensis, 367.
sanguinis hominis, 367, 368,

369, 379,381,382,383.
Smithii, 221,237, 238,258.
tercbra, 358.
Filaria, Relationship of Mosquito to,

368.
Filarial, 368, 370, 371, 372, 373,

377-387.
Foraminifcra, 444.
Formica, 141, 143.

rufa, 418.
Fossil Walrus, found in the Red Crag
of Suffolk, On the Tusks of the,
Prof. E. Ray Lankester, 213.
Fryeria, 125, 137.

variabilis, 125, 137, 140.

Gammarus, 402.
Gasterocoma, 9.
Gastrodiscus, 242.

Sonsinonis, 241.
Gastrophilus, 224, 360.

elephantis, 224, 246, 247, 249,
255.

rhinocerotis, 248.
Geotriton(Desmognathus) fuscus, 208.
Gephyrseans (Echiurid), 469, 471.
Glaucus, 123.
Glenotremites, 8, 14, 73, 74, 75, 90.

conoideus, 73.

Jolfingia, 469, 470, 471, 473.

Maclntoshii, 469, 473, 474.
3onepteryx, 333, (prehensors) 333.
Clorinde, 333.
Leachiana, 333.
Ehamni, 333.
Goniaster, 82.
Goniodoris Tyroni, 129.
Gordius, 230, 381.
Gosse, Philip Henry, On the Clas-
ping-organs ancillary to Genera-
tion in certain Groups of the
Lepidoptera, 265.
Graff, Prof. L. von, Note on a new

Myzostoma, 435.
Great Rods in Insect Eye, 406.
Gryllus, 147, 148.
Gynmodoris, (footnote) 134.
Gymnura, 261, 262, 264.

rafflesii, 260, 261, 262, 264.
Gyrinophilus porphyriticus, 208.

Haddon, A. C, On the Extinct Land-
Tortoises of Mauritius and Rodri-
guez, 155.
Hoematomyzus, 249.

elephantis, 224, 249.
longirostris, 249.
Haematopinus, 249.
Hag-fish, 179.
Hair, thatch of, 271, 313, 314, 326,

329, 331.
Hamingia, 469.
Harmothoe areolata, 351, 354.
Harpe, armaturo of, 272, 315, 316,
318, 319, 324, 325, 326, 327 ;
base of, 272, 294, 296, 319, 325,
328, 337 ; character of, 271 ;
disguises of, 329, 330, 331 ; func-
tion of, 273 ; glitter of, 272, 299,
315, 323, 330, 331 ; type-forms
of, 280-282.
Hatteria, (footnote) 193.
Hawk-Moth, pupa of, 420.
Hebomoia, (prehensors) 333.

Glaucippe, 333.
Helmintha, 224.

Hemerophila perfumaria, 398, 401.
Hemiptera, 141-154, 417.
Hemiptera-Heteroptera, 249.
Hermadion fragile, 352.
Herold on Pieris, 265.
Hertha mystica, 8, 9.
Hesperiadse, 266, (footnote) 298.
Heteronereis, 352.

INDEX.

Heteroptera, 249.
Hexabranchus, 133.
Holopus, 443.
Holothuria?, 125.
Homo sapiens, 260, 264.
Homopus, 256.

elephantis, 224, 250.
Hymenoptera, 142, 143, 144, 151,

399, 400, 101, 404, 408, 409, 416,

425.
Hyocrinus, 109.
Hyperia, 405.

Ichneumon, 423.

Ichthyopsida, 206.

Idolocoris elephantis, 249, 256.

Inarticulata, 5.

Insecta, 224, 416.

Insectivora, 262.

Isis, 2.

Isocardia cor, 13

Isocrinus, 109.

Isopods, eyes of, 396, 414, 416.

Kallah-nimok, 251.

Lamprey and Hag-fish, 179.
Land-Tortoises of Mauritius and
Rodriguez, On the Extinct, by A.
C. Haddon, 155.
Lankester, E. Ray, On the Tusks of
the Fossil Walrus, found in the
Red Crag of Suffolk, 213.

Golfingia Maclntoshii, a new
Sipunculid from the Coast of
Scotland, 469.
Lasioeampidse, 425.
Lasius, 141.

flavus, 141, (prothorax) 143,

144, 148, 151, 153. 154.
fuliginosus, 151.
Lemgum, 252.
Leon-Dufour, on urinary organs in

Insects, 287.
Lepidonotus, 347.
clava, 347.
clavus, 347.
squamatus, 347, 348.
Lepidoptera, 265-345, 389-420;
scales in, 269, 273, 280, 299, 305,
307, 311, 319.
Lepidoptera, On the Clasping-organs
anciUary to Generation in certain
Groups of, P. H. Gosse, 265.
Leptoeircus (footnote), 334,
Curius, 267.

485

Leptodora hyalina, 143.

Lethrea geognostica, (footnote) 14,

(footnote) 20.
Libellula depressa, 419.
Linmrea truncatula, 250.
Limulus, On the Testis of, by W. B.

S. Benham, 363.
Limulus, 363, 365-366; course of
spermatic network, 364 ; external
aperture, 364; lateral network,
364 ; median network, 364 ; ob-
servations, 365 ; spermatozoa, 365 ;
sperm-sacs, 365.
Liparida?, 425.
Liparis, 425.
Lipsiffi, (footnote) 1.
Lissotriton punctatus, 207.
Loris gracilis, 260.
Lowne, B. T., On the Compound Vision
and the Morphology of the Eye in
Insects, 389.
Loxodonta africana, 230, 231, 247,

256.
Lubbock, Sir J., On the Anatomy
of Ants, 141.

Mammalia, 197, 259.

Manipulation, 268.

Manson, P., The Metamorphosis of
Filaria sanguinis hominis in the
Mosquito, 367.

Marana, 425.

Marsupites, 14.

Masuri, 238, 251, 252, 254.

Mauritius, Tortoises of, (Haddon),
156.

Meconium, foreign, 269, 307 ; na-
ture of, 288, 311.

Medusa?, 411.

Megaderma lyra, 261.

Melokmtha, 145, 146, 149.

Menobranchus, 167, 173, 178, 188,

189, 191, 192, 193, 195, 196, 206,
208.

lateralis, 191.
Menopome, (footnote) 175, 188, 189,

190, 196, 203, 206, 209.
Sieboldia, 188, 196, 206.

Mermis, 381.

Mesozoic Crinoids, 5.

Metacrinus, On three new species of,

P. H. Carpenter, 435.
Metacrinus, 435, 436, 437, 438, 439.

440, 441, 443.

angulatus, 438, 439, 448.

486

INDEX.

Metacrinus

cingulatus, 439, 443.

costatus, 444.

interruptus, 438, 439, 443, 444.

Moseleyi, 437.

Murrayi, 442,

nobiUs, 439, 441, 442.

rotundus, 435-440, 442, 443,
444, 446.

Stewarti, 436, 443, 444, 44G.

superbus, 435, 440, 441. 442,
444, 446.

Wyvillii, 443.
Metamorphosis of Filaria sanguinis
hominis in the mosquito, by P.
Manson, 367.
Mollusca, 411.
Molpadida>, 38.
Monofcreme, I!
Morpho. 345. (prehensors) 334.

Epistrophis, 334.

Menelaus, 334.
Morphology of the skull in the Am-
phibia Urodela, ou the, Prof. W.
K. Parker, 165.
Morphology of the eyes of Arthro-
pods, 414-417.
Morphology of the Eye in insects

(Lowne), 389.
Mosquito, relationship of, to Falaria,

367, 368, 370 386.
Muridte, 262.
Mus, 260.

decurnanus, 260, 261.

natalensis, 203.
Musca vomitoria. 396.
Muscida?, 399, 401, 414,417.
Museulus, 259.

depressor mandibular, 259, 262.

digastricus, 259.
Mussoorie, 238, 251, 252.
Myoxidae, 2 12.

Myoxus avellanarius, 260, 261, 263.
Myriapoda, 416.

Myrmicaruginodis, 148,149,151,153.
Mysis, 403.
ilyzostoma, note on a new, Prof. L.

von Graff, 435.
Myzostoma, 435, 444, 445.

cirriferum, 445.

cirripedium, 440, 444, 446,

'Wyville-Thonisoni, 444.

Nematocerous Diptera, 400,
Nematoda, 224, 252.

Nematodes, 231, 251, 252, 253.
Neuron, anatomy and functions of,

(Lowne), 407-111.
Neuroptera, 399, 408-420.
Nisoniades, 266.
Noctuid Moth, 408-420.
Notidanus, 206.
Notocotylus, 24 1 .
Notodon (footnote), 400.
Notonecta, 394. 398, 399, 405.
Nbtophthalnms, 177. L83, L85, 186,

190, 192, (footnote) 193, 199, 228.
viridescens, 177, 180, 208.
Nudibranchiata, 123, 126, 127.
Nudibranchiate Mollusca from the

Eastern Seas, on some newspei ies,

bj C Collingwood, 123.

list of Genera and Species, L25.
Xumenes, 425.

(Estrus elephantis, 240, 247.
Oligoneuria rhenana, 288.
Onchnesoma, 473.
Onychodactylus, (skull)]--.
Ophiactis virens, (footnote) 36, 57.
Ophiocrinus, 16.
Ophiura texturata. 37.
Ophiurida?, 1. 4. 30. 55.
Ophiuroida?. 2.
Ophiuroidea, 2.

Organs, general description of. 270 ;

valves, 270 : harpes, 271 : uncus,

273 ; scaphium, 270 ; penis, 277.

Ornithoptera. 266, 207, 270, 271,

273,274,281,282,286, 2-7. 293,

294, 299, 311, 331, (footnote)

336, 338.

Amphrisius, (footnote) 200, 207,

281, 290, 2 13,330.
Amphrysus, (footnote) 207,
(footnote). 271, 274. 276,
281, 285, 287, 292,339.
Arruana, (footnote) 271, 274,

275. 281-284, 2-7. 339.
Brookeana, (footnote) 271, 275,

281, 291,292, 339.
Darsius, 27-. 281, (penis of)

286, 287, 2! 0-292, 339.
Haliphron, 270, (footnote) 271,
277. 27-. 281, 284, (Sperma-
tozoa of) 285, 286, 2-7, 289-
292, 294,316,335,336,339.
Heliacon. 281, 290, 339.
Heliaconoides, 2S1, 290, 339.
Priamus, 299.

Ornithoptera

Pronomus, 281, 283, 284, 339.
ltemus, (footnote) 271, 272,
(footnote) 276, 277, 281, 291,
293, (footnote) 304, (footnote)
305, 335-340.
Ehadamanthus, 267, 269,
(foreign matters in) 269,
( footnote) 271, 281, 285, 287,
289, 290, 291, 293,339.
Richmond'a, 278, 281, 283,
2-4. 286, 2-7, 339.
Ornithoptera;, 27(5. 27S, 319.
Orthoptera, 404, 417.
Osseous fishes, 168, 1-2.

Pachydermata, 256.
icrinoids, 439.
Papilio. 266, 207. 270, 272, 273,
(footnote) 27:', 281, 332, 334,
337, 338.

Agamemnon, 270, 274, 276,
27-. 281, 286,314,337,338,
342.
Agavus, 274. 278, 280, 281,

301, (footnote) 302,441.
Agenor, 295.
Agesilaus, 272, 274, 278, 282,

326, 344.

Anchisiades, 270, 279, 281,315,

318, 319,338, 342.
Androgeos, 295, 310, 311.
Antenor, 274. 282, 331, 344.
Archesilaus, 274, 282, 324, 327,

343.
Arcturus, 275, 270, 281, 303,

(footnote) 304, 308,319,341.
Arruana, 274, 275.
Ascalaphus, 272, 2S0, 281, 300,

301, 337, 340.
Axion, 269, (footnote) 271, 272,

281,311, 342.
Bathycles, 275, 270, 277, 278,

2-1. 313,342.
Bromius, 208, (footnote) 271,

274, 277. 281, 308, 320,

(footnote) 328, 343.
Brutus, 2S1, 306, 341.
Chaon. 281, 300.
Children®, 272, 274, 2S1, 316,

317, 342.
Codrus, 274, 278, 282, 2S6, 321,

327, 344.

Coon, (footnote) 266, 267, 277,
278, 323.

'apilio

Deiphontes, 274, 277, 281, 297,

300, 301, 3-10.
Demoleus, 268, 274, 275, 231,

299, 300, (footnote) 305, 312,

314, 322, 342.
Diphilus, 271, 274, 282, 329,

(aberrance of), 329, 330, 331,

344.
Doubledayi, 207, 278, 281, 284,

323, 343.
Erechtheus, 272, 274, 275, 276,

278, 231, 296, 299, 340.
Ergeteles, 2S1, 319, 343.
Erithonius, 268, 270, 272, 275,

278, 279, 281, 286, 314, 315,

342.
Euryleon, 270, 272, 281, 317,

318, 343.
Gigon (spermatozoa in), 288.
Hector, 271, 274, 273, 280,
282,286,287,301,302,322,

331, 332, 344.
Helenus, (footnote) 266, 274,
278, 279, 280, 281, 300, 309,
341.
Hesperus, 270, 274, 277, 278,

281, 312, 342.
Homerus. 274, 275, 231, 307,

322, 341.
Latreillii, 281, 321, 343.
Liris, (footnote) 200, 267, 330.
Lyeidas, 270, 274, 270, 277,

281, 322, 343.
Lyeophron, 310, 311.
Lysander, 281, 317, 319, 343.
Macedon, 272, 274-27-, 281,

319, 338, 343.
Machaon, (footnote) 296, 274,
275, (footnote) 276, 281,302,
(footnote) 304, 341.
Mavo, 272. 274. 275. 278, 281,
295, 297, 303, 309, 337, 340.
Memnon, (footnote) 266, 272.
275, 278, 281, 294, (continen-
tal forms of) 295, 300. 301,
337, 345.
Menestheus, 274, 281, 301,

341.
Merope, 274-277, 2S1, 304,
306-308,313,314,330,341.

Nephelus, 231. 300.
Niieus,268,275,278,282,308,

320, 32-, 314.
Oebalus, (footnote) 310.

INDEX.

Papilio

Pammon, (footnote) 266, 272,

275. 278, 281, 301,341.
Parmatus, 272, 282, 327, 344.
Phoresis, 268, 281,308,341.
Podaliriti*. 272, 275-278, 2-1,

286, 320,321, 343.
Polices. 270, 2-i'. 325, 344.
Polycaou, 278, 231, 310, 311,

342.
Polydanias. 279.
Polydorus, (footnote) 266, 271,
274, 282, 2^4, 302, 329-332,
344.
Polvmnestor, 272, 275, 281,

295, 297, 299, 340.
Protenor, 281, 297, 299,340.
Ehetenor, 272, 274-278, 281,

297, 299, 34u.
Ehodifer. 274, 275, 277, 27-,

2-1. 323, 343.
Schmeltzi, 231. 298, 340.
Semperi, (footnote) 271, 278,

281, 322, 343.
Sesostris, 278, 281, 316-31-,

342.
Thersites. 310. 311.
Thoas, 275. 270.281,308,309,

311, 341.
Turnus, 274, 275, 277, 2-1.

309, 342.
Ucalegon, 270, 271, 277, 2-2.

325, 344.
Ulysses, 272, 275. 278, 231.

307, 308, 310, 341.
Yertunimis, 272, 274, 275, 277,

278, 2-1. 317.342.343.
Zalmoxis, 269, (foreign matteis
in) 209, 270. (footnote) 271,
272,274,276,278,282,324,

343.
Zolicaon, (footnote) 304.
Papiliones, 281,327, 334.
Papilionid«. 205. 271, (footnote 1 299,

(footnote) 330, 334.
Papilionides, 330.
Parasites of Elephants, by T. Spencer

Cobbold. 21':',.
Parasites, table of, 234.
Parker, W. K., On the morphology

of the skull in the Amphibia

"Crodela. 165.
Parnassius Apollo, 267, (footnote)

268, 274.
Pelobates fuscus, 207.

487

Penis (of Lepidoptera), axial move-
ment, 273 : base, 337 ; change in.
287 : (footnote) 300,313, 323, 330 ;
characters of, 277, 330 ; forms ot.
278, 1-7,321,324.323,336; large
size of, 322 ; Owen on, 299, posi-
tion of, 277 ; protrusion of, 277,
313, 316, 319, 320, 323, 324:
sheath of, 297, 313, 314, 315, 337 :
in Ornithoptera Haliphron, 285,
294 ; in O. Itemus, 336 : in Pa-
pilio Codrus, 323 : in P. Lysander,
318 ; in P. Ehodifer, 324 ; in P.
Semperi, 322.
PentacrinidiB, 14, 438, 439, 440, 481 1.
Pentaerimtes, 5, 6.
Pentacrinoids. 475.
Pentacrinus, 6, 8, (footnote) 9, 10.
i'i.i. 21, 22. (footnote) 23, 53-57,
(footnote) 64, 67. 68, 71. 73. 75, 90,
104-109, 435, 43-, 440. 441, 443.
asteria, 23. 68, 105, 109.
asterius, 439, 441. 442.
briareus, 105, 109.
caput-Medusa?, (footnote) 8. 23.

68, (footnote) 442.
Maclearanus, 441.
Hulleri, 23, 125. 441.
scalaris, 109.
subangularis, 105.
Wyville-Thomsoni, 57. 105,
441, 443.
Pentastomes, 360.
Perennibranehs, 173, 179, 191. 194.

197, 100.
Petalostoma, 473.
Pezophaps solitaria, 15, .
Phalangium, 411-.
Phanogenia, 16, 20, 28. 43. 100.

typica. 20.
Phantasmocoris, 249.
Fhanngobolus africanus. 255. 2o6.
Phascolion. 472.
Phascolosoma Hanseni, 472.
Pheidole megaeephala, 14-.
Phoca communis, 261.
Pholadomya, 220.
Phronomia. 4 15.
Phvcidx\ 425.
Phyllidia, 125, 136.

pustulosa, 136.

spectabilis, 125, 136, 140.

verruculosa. 136.
PhyUidiadte, 125. 136.
Phyllidiella pustulosa. 125, 130.

483

INDEX.

Pbymosoma, 470, 472.

Ruppellii, 472.
Picumnus minutus, (footnote) 181,

L93.
Pieridae, Herold on, 265, (footnote)

271, 277,332.
Pieris, 265, (footnote) 279, 332, (pre-
hensors) 332.

Autodice, 332.

Automate, 332.

Brasfricae, 265, 276, 332, 405.

Rapae, 332.
Pipa, 170, 2o7.
Plethodon glutinosus, 208.
Pleurodeles Waltlii, (footnote) 209.
Polybranchiate, 125.
Polyoeridae, 133.
Polygordius, 353.

Polynoe, 347, 348, 349, 350, 351,
352, 354,355.

areolata, 351, 352, 354, 355.

clava, 347-352, 354, 1555.

grubiana, 347, 348.

lunulata, 353.

pellucida, (footnote) 353.

praeclara, (footnote) 353.

spinifera, 352, 354, 1555, 356.

squamata, 347-352, 355.

squamatus, 347.

Wahlii, (footnote) 353.

(Antinoe) praeclara, (footnote)
353.

( ) Wahlii, (footnote) 353.

(Harmothoe) areolata, 351, 354.
Polynoina, 350, 351, 352, 356.
Polypterus, 187.
Poreellio, eyes of, 415.
Prehensors, material, 268 ; use for
diagnosis, 26S ; variety of, 207,
273.
Primates, 259, 202.
Proteus, 178, 188, 189, 191 -196,
206, 209.

anguinus, 188, 208.
Pseudis paradoxa, 193.
Pseudophryne Bibronii, 207.
Psittacus, (footnote) 193.
Pterocoma, 8, 9.

Pterophorus pentadactylus, 391, 417.
Pteropus, 261.

edulis, 261.

medius, 261.
Pyrula, 125, 220.

Quadrumana, 263.

Rana teroporaria, 172, 173.
Ranodon, 209.

sibiricus, 208.
Relationship of Mosquito to Falaria,

368.
Rcptilia, 259.
Retinula, 391, 397.
Rhinoceros, 248.

bicornis, 248.

indicus, 248.

Bimus, 2 18.
Rhizocrinue, (footnote) 23, 57, 68,
(footnote) 73, (footnote) 106, 107,
109.

lofotensis(footnote), 68,107, 10'.).

Ftawsonii, 109.
Rhyacophila, (footnote) 279.
Rodentia, 259, 263.

Saecocoma, 8, 1 I.
Salamandra, L65, 17-, L79, 182.
atra, 208.

maculosa, 165, 17!). 188, 192.
(embryo) 17", (skull) L72, 180,
L88, 192.
Salamandrina perspieillata, 208.
Salmo, -117. 150, 162.
alpinus, 462.
fario, (footnote) 450.
fontinalis, 462, 465.
levenensis, 402.
salar, 447, (footnote) 450, 455,

459, 461, 462, 466, 408.
trutta, (footnote) 450, 462.
Salmon, 447-468 ; arrest of growth

in young, 465, 466.
SalmonidaB, (footnote) 457, 459, 461,

463, 167.
S dmonoids, 450, 453, 462, 465.
Salticus, 415.

scenicus, 404, 415, 418.
Saprolegnia ferax, 457.
Sargassum, 138.

bacciferum, 138.
branch of, -with coil of spawn of
Scylla a pi lagica, 138.
Sauropsida, 194

Scales in Lepidoptera, foreign. 269,
273, 280, 299, 305, 307, 31 1,319.
Scapbium, bristles of, 276, 285, 289,
297, 310, 316, 320, 325; cha-
racters of, 275 ; cheeks of, 335 ;
colour of, 276, 289, 305, 319, 325,
326, 330 ; forms of, 275, 313 ;
function of, 276, 305, 306 ; keel

of, 283, 289, 292, 301, 309; in

Ornithoptera Remus, 335 ; in

Papilio Agamemnon, 337 ; in P.

Anchisiades, 338 ; in P. Asca-

laphus, 337 ; in P. Codrus, 328 ;

in P. Macedon, 338; in Pieridse,

&c, 332 ; pentagon, 335 ; rami,

276, 296, 297, 300, 304, 313;

resemblance to a boat. 275,206,304,

to a jaw, 275, 295, 304 ; teeth of,

275, 296, 304, 305, 313, 330, 331.

Sclerostoma, 224, 229, 230, 256.

armatum, 229.

olatlir.it urn, 229, 230.

sipunculiforme, 224, 226, 228,

230, 231, 232,256,257.
tetracanthus, 229.
Scorpions, 365—4 L6.
Scudder and Burgess, on Xisoniades,
266 : on asymmetry in ditto, (foot-
note) 298.
Scyllsea, 123, 137, 138.
marmorata, L38.
pelagica, 123, 125, 137, 138,
(spawn of, attached to a
branch of Sargassum) 138, 140.
viridis, 138.
ScyllaBidas, 137.
Sea-Stars, 1, 3, 5.
Sea-Trout, 449, 459.
Seironota, 168, 194.

perspieillata, 207, 208.
Sericinus, (footnote) 334.
Shotee, 252.
Sialis, larva, eyes of, 415.
Sieboldia, 188, 200, 209.
Sigalion, 354.

Simondsia paradoxa, and on its

Probable Affinity with Sphoeru-

laria bombi, by T. S. Cobbold, 357.

Simondsia, 357, 358, 359, 360, 361.

paradoxa, 357, 358, 359, 361.
Siphneus, 262.

Sipunculid worms, 469, 470, 471.
Sipuneulidean Gephyrajans, 471.
Sipunculoidea, 471.

Key to genera of, 472.
Sipunculus, 469, 470, 472.
australis, 472.
edulis, 472.
nudus, 471.
punetatus, 469.
Siredon, 178.

Siren, 188-190, (footnote) 193, 196,
206, 209.

INDEX.

4B9

Siren

lacertina,188(footnote),193,2'i;

•27?,.

Smerinthus populi, 396, 39
Solanocrinites, 5, 6, 9.
oostatus, 6.
Jsegeri, 6.
scrobiculatus, 6.
Solanocrinus, 8, 9, 14, 10, 71, 73,
85, 104, 105, 100, 109.
oostatus, 104, 105.
Jiegeri, 104, 105, 106.
scrobiculatus, (footnote) 85, 104,
105.
Soorti, 252, 253, 254.
Spalax, 262.

Spelerpes, (footnote) 175, 187,191,
(footnote) 193,194, 199,203 205.
fuscus, 208.

rubra, (skulls) 191. 196, 19s.
salmonea, (skulls) 191, 195, 196.
Spermatozoa, Gladstone on. 287; in
Neuroptera, 228: in penis,
285, 286, 288, 316, 318.
Spba?rularia, 358, 359.

bornbi, 358.
Sphierulariae, 358.
Spotted Salamander, 189.
Stella. 2.

barbata, 2.
chinensis, 2, 5.
decacnemus rosacea, 4, 5.
fimbriata, 2.
rosacea, 2.

(Decacnemus) rosacea, 2.
Stellerites, 5.
Stcnodoris rubra, 134.
Stilasteritte, 5, 6.
Strongylus, 224, 225.

clathratus, 224, 230, 231, 232,

234, 257.
elephanti, 225.

falcifer, 224. 234, 235, 257,258.
foliatus,224,233,234,237,257.
Sus scrofa, var. domestica, 359, 361.
Syllidoe, 351.

Symbiotes elepbantis, 250.
Syrphidse, 393.
Syrphus, 420.

ribesii, 419, 420.

Teinopalpus, (footnote) 334.
Teleostean Fishes, 182.
Telephorus, 394, 409, 410.
Terias, 333, (prehensors) 333.

Hecabe, 333.
Tessellata, 14, 21, 22, 24.
Testudo, 156.

indica, 156.

inepta, 156, 157, 158. 150, L62,
163.

, var. Boutoni, 157.

Troglodytes, 261, 202.

niger, 261, 264.
Trogon, (footnote) 103.
Tupaia, 262, 264.

Ellioti, 260, 261, 263.
Tylosoma, 473.
Typhis, 390, 405.
Tyroglyplrus siro, 250.

Tabanida?, 414.

Tachina onchestus, 425, 426.

Taeniae, 360.

Taricha, 187.

torosa (skull), 186, 187, 208.

SECOND SERIES. — ZOOLOGY, VOL. II

leptocnemis, 156, 159.
triserrata, 156, 157, 158, 159,

160, 161, 162, 103.
Yosmteri, 150, 159, L63.
Tethys, 127.

Teuthophagous Whales, 220.

Thais Hypsipyle, 267.

Tipula, 394, 398, 404, 405. 407,

409, 410,416,418 410.

oleracea, 419, 420.

Tortoises of Mauritius and Rodriguez,

On the Extinct Land, by A. C.

Haddon, 155, 156.

Trachearia, 246.

Trematoda, 224, 256.

Trematodes, 254.

Trevelyana, 124, 125, 134.

felis, 125, 134, 140.

rubra, 134.

(Stcnodoris) rubra, 134.

Tnchecodon, 213, 214, 215.

Huxleyi, 213, 214, 215, 216.
Koninckii, 214, 215.
Triehecus, 213, 214, 215, 216, 217.
Huxleyi, 213, 214, 215, (tusks)

216, 2 is.
(Trichecodon) Huxleyi, -15, 2 10.
rosmarus, 216, (footnote) 217.
Trichina, 367.
Trichoccphalus, 357.
Trichoptera, McLachlan on, 227.
Trichuris, 357.
Triopa, 133.

Principis-Wallice, 125, 133,
140.
Triphsena pronuba, 398, 41 S.
Triton, ISO, 206.

cristatus, ISO, 192, 207.
platycepbalus, 208.
subcristotus, 208.
torosus, 208.
virideseens, 208.
(Xotophthalmus) virideseens,
208.

Uncus, absent, 313, 317, 321,328:
characters of, 273 ; forms of. 274.
329, 330 ; in Ornithoptera Remus,
335; minute, 320: resemblances t
274.
Uniformity, individual, (footnote)

289, 320.
Urine, in Insects, 2S7 : in Serpents,

288.
Urodela, (footnote) 165, 107, 183.
Urodele, 173, 177,206.
Urodeles, 171, 172, 173, 190, L92,
(footnote) 193, 196, 203, 206,
207, 208.

Valves, asymmetry in, 298 ; cha-
racters of, 270 ; disguises of, 320,
330, 331 ; hinge-knobs, 337 ;
Wallace on, 271.
Vanessa, 405.

Variations, Purpose of Structural, 270.
Variety in Structure, 267, 273 :
object of, 270; specific, (footnote)
289, 300,311, 317, 320.
Verruca, 444.
Ycspa, 358.

germanica, 418.
rata, 358.
vulgaris, 358.
Yoluta, 220.

Wallace, on Ornithoptera, 271: on

Papilio Memnon, 205.
Walrus-tusks, (growth) 219, (recent

and fossil) 217.
Walsingham, Lord, Contributions to

the Knowledge of the Genus

Anaphe, 421.
White on Rhopalocera, 205, 207 ; on

Papilio Podalirius, 321.

Xylocopa, 41S.
Xylophasia polyodou, 418.

Ziphioids, 220.
Zophodiopsis hyainella, 425.

76

1' 11 I N T ED I! V T A V I. C) K A N I) 1 1! I N C I S,

RKIi LION CO! Itl . II, Kl I M m.l. i

II I I ll
3 2044

II

06 309 313

Date Due

Dte- ~W3